U.S. patent number 6,077,708 [Application Number 09/160,457] was granted by the patent office on 2000-06-20 for method of determining progenitor cell content of a hematopoietic cell culture.
Invention is credited to Paul C. Collins, William M. Miller, E. Terry Papoutsakis.
United States Patent |
6,077,708 |
Collins , et al. |
June 20, 2000 |
**Please see images for:
( Certificate of Correction ) ** |
Method of determining progenitor cell content of a hematopoietic
cell culture
Abstract
The invention provides a method of determining the content of
progenitor cells in a hematopoietic cell culture. The method
comprises measuring one or more metabolic parameters of the
culture, such as oxygen consumption, glucose consumption and/or
lactate production, and using the measured parameter(s) to
determine the content of progenitor cells.
Inventors: |
Collins; Paul C. (Bridgewater,
NJ), Papoutsakis; E. Terry (Glenview, IL), Miller;
William M. (Evanston, IL) |
Family
ID: |
27368311 |
Appl.
No.: |
09/160,457 |
Filed: |
September 24, 1998 |
Related U.S. Patent Documents
|
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
Issue Date |
|
|
116853 |
Jul 16, 1998 |
|
|
|
|
Current U.S.
Class: |
435/375; 435/14;
435/325; 435/372.2; 435/372.3; 436/14; 436/62 |
Current CPC
Class: |
G01N
33/5094 (20130101); Y10T 436/104998 (20150115) |
Current International
Class: |
G01N
33/50 (20060101); C12N 005/00 (); C12N
005/08 () |
Field of
Search: |
;435/325,372,372.2,372.3,375,289.1,14 ;436/14,62 |
Other References
Bainton et al., J. Exp. Med., 134, 907-934 (1971). .
Beck et al., Cancer Res., 12, 823-828 (1952). .
Beck et al., Cancer Res., 12:818-822 (1952). .
Beck, J. Biol. Chem., 232:251-270(1958). .
Bender et al., J. Hematotherapy, 1:329-341 (1992). .
Bertolini et al., Blood, 89:2679-2688 (1997). .
Bird et al. Cancer, 1009-1014 (1951). .
Brugger et al., New Engl. J. Med., 333:283-287 (1995). .
Cannistra et al., J. Biol. Chem., 265(21):12656-12663 (1990). .
Cannistra et al., Proc. Nat'l Acad. Sci., 87:93-97 (1990). .
Cline, in Formation and Destruction of Blood Cells, pp. 222-239
(Greenwalt and Jamieson eds., 1970). .
Collins et al., Glucose and Lactate Metabolic Rates Predict Numbers
of Colony-Forming Cells in Hematopoietic Culture, Abstract 2417,
pub. Nov. 15, 1996. .
Collins et al., Biotechnol. Bioeng., 55:693-700 (1997). .
Collins et al., Biotechnol. Prog., 14:466-472 (1998). .
Collins et al., Biotechnol. Bioeng., 59:534-43 (1998). .
DiPersio et al., J. Biol. Chem., 263(4):1834-41 (1988). .
Gesinski et al., Australian Journal of Biological Sciences,
21:1319-1324 (1968). .
Gore et al., Exp. Hematol., 23:413-421 (1995). .
Hamilton et al., Biochem. Biophys. Res. Commun., 138:445-454
(1986). .
Hamilton et al., J. Cell. Physiol., 134:405-412 (1988). .
Haylock et al., in: Hematopoietic stem cells: biology and
therapeutic applications (D. Levitt et al., eds., Marcel Dekker,
Inc., New York, pp. 491-517 (1995). .
Hedeskov, Biochem. J., 110:373-380 (1968). .
Hume et al., J. Natl. Cancer Inst., 62:3-8 (1979). .
Kester et al., Arch. Biochem. Biophys., 183:700-709 (1977). .
Koller et al., Exp. Hematol., 20:264-270 (1992). .
Koller et al., Biotechnol. Bioeng., 50:505-513 (1996). .
Koller et al., Ann. New York Acad. Sci., 665:105-116 (1992). .
Koller et al., Blood, 80:403-411 (1992). .
Laluppa et al., Exp. Hematol., 26:835-843 (1998). .
Laluppa, J. Biomed. Mat. Res., 36:347-359 (1997). .
Lanks et al., J. Cell. Physiol., 135:151-155 (1988). .
Lutton et al., Experientia, 28:850 (1972). .
McAdams et al., Trends Biotechnol., 14:341-349 (1996). .
McAdams et al., Trends Biotechnol., 14:388-396 (1996). .
Miller et al., J. Cell. Physiol., 132:524-530 (1987). .
Miller et al., Bioprocess Engineering 3:103-111 (1988). .
Newsholme, et al., Biosci. Rep., 5:393-400 (1985). .
Olander, American Journal of Physiology, 222:45-48 (1972). .
Peng et al., Annals of Biomedical Engineering, 24:373-381 (1996).
.
Pierson et al., J. Hematotherapy, 5:475-483 (1996). .
Rogers et al., Ann. Hum. Genet., 43:213-226 (1980). .
Sand et al., Blood, 50:337-346 (1977). .
Sardonini and Wu, Biotechnol. Prog., 9:131-137 (1993). .
Spielholz et al., Blood, 85:973-980 (1995). .
Testa et al., Blood, 88(9):3391-3406 (1996). .
Traycoff et al., Exp Hematol, 22:1264-1272 (1994). .
Whetton et al., EMBO J., 3:409-413 (1984). .
Whetton, et al., J. Cell Sci., 84:93-104 (1986). .
Williams et al., Blood., 87:1687-1691 (1996). .
Zandstra et al., Bio/Technology, 12:909-914 (1994). .
Zandstra et al., Biotechnol. Bioeng., 54:58-66 (1997). .
Zhou et al., Biotechnol. Bioeng., 44:170-177 (1994)..
|
Primary Examiner: Mertz; Prema
Assistant Examiner: Hamud; Fozia
Attorney, Agent or Firm: Ross; Sheridan
Government Interests
This invention was made with government support under National
Science Foundation Grant BES-9410751 and National Institutes Of
Health Predoctoral Biotechnology Training Grant GM08449. The
government may have certain rights in the invention.
Parent Case Text
This application is a continuation-in-part of application No.
09/116,853 filed Jul. 16, 1998 now abandoned. Also, benefit of
provisional applications No. 60/052,958, filed Jul. 18, 1997, and
60/059,811, filed Sep. 24, 1997, is hereby claimed.
Claims
We claim:
1. A method of determining the content of progenitor cells in a
hematopoietic cell culture, the method comprising:
culturing a reference culture of hematopoietic cells under selected
conditions, the selected conditions comprising employing a selected
feeding protocol;
measuring one or more metabolic parameters selected from the group
consisting of glucose consumption, lactate production and oxygen
consumption of the reference culture at a plurality of selected
times;
determining the content of progenitor cells in the reference
culture at each of the selected times;
culturing an experimental culture of hematopoietic cells under
essentially the same conditions used to culture the reference
culture;
measuring the same one or more metabolic parameters for the
experimental culture at one or more selected time(s) as measured
for the reference culture; and
comparing the one or more metabolic parameters measured for the
experimental culture with those measured for the reference culture
to determine the content of progenitor cells in the experimental
culture at the selected time(s).
2. The method of claim 1 wherein the selected conditions comprise
culturing bone marrow mononuclear cells, cord blood mononuclear
cells, or peripheral blood mononuclear cells.
3. The method of claim 1 wherein the selected conditions comprise
culturing a cell population from bone marrow, cord blood or
peripheral blood that is enriched in CD34.sup.+ cells.
4. The method of claim 1 wherein the selected conditions comprise
using a combination of cytokines that causes expansion of total
cells, progenitor cells, post-progenitor cells or combinations
thereof.
5. The method of claim 1 wherein the selected conditions comprise
using at least one cytokine from each of the following groups:
i. cytokines acting on primitive hematopoietic cells; and
ii. cytokines acting on a wide array of progenitor cells.
6. The method of claim 5 wherein the selected conditions further
comprise using at least one additional cytokine from the following
additional group:
iii. cytokines acting on lineage-restricted hematopoietic
cells.
7. The method of claim 1 wherein the selected conditions comprise
using at least one cytokine from each of the following groups:
i. stem cell factor or Flt3 ligand;
ii. interleukin-3, interleukin-6, or granulocyte-macrophage colony
stimulating factor; and
iii. granulocyte colony stimulation factor or erythropoietin.
8. The method of claim 1 wherein the selected conditions comprise
using serum-free culture medium.
9. The method of claim 1 wherein the selected conditions comprise
using serum-containing culture medium.
10. The method of claim 1 wherein the selected conditions comprise
stirring the culture.
11. The method of claim 10 wherein the stirred culture is performed
in a bioreactor.
12. The method of claim 1 wherein the selected conditions comprise
not stirring the culture.
13. The method of claim 1 wherein the selected conditions comprise
using a feeding protocol which comprises adjusting the cell density
of the culture at least every other day to a selected level so as
to optimize the production of cells.
14. The method of claim 13 wherein the cell density is adjusted
daily to about 1.5-2.0.times.10.sup.6 cells/ml.
15. The method of claim 1 wherein the selected conditions comprise
using a feeding protocol which comprises replacing about 50% of the
culture medium, with cell retention, every other day beginning
about 4 days after initiation of the culture.
16. A method of determining colony-forming cell content of a
hematopoietic cell culture, the method comprising:
culturing a reference culture of hematopoietic cells under selected
conditions, the selected conditions comprising using a feeding
protocol that limits the production of post-progenitor cells that
have high glucose consumption and lactate production;
measuring the total glucose consumption or lactate production per
unit volume of the reference culture at a plurality of selected
times;
measuring the density of nucleated cells (X.sub.i) present in the
reference culture at each of the selected times;
using the measured X.sub.i and the measured glucose consumption or
lactate production to calculate q.sub.gluc or q.sub.lac at each of
the selected times using equations (1) and (2) set forth below:
##EQU17## wherein: Q.sub.i is the total glucose consumed or lactate
produced (.mu.mole/ml/hr) at time t.sub.i ;
t.sub.i is any time point;
t.sub.i+1 is a time point after time point t.sub.i ;
t.sub.i-1 is a time point before time point t.sub.i ;
slope.sub.b is the first order backward slope from time t.sub.i-1
to time t.sub.i of the total glucose consumption or lactate
production curve calculated by dividing the point-to-point glucose
consumption or lactate production differences by the point-to-point
differences in time; and
slope.sub.f is the first order forward slope from time t.sub.i to
time t.sub.i+1 of the total glucose consumption or lactate
production curve calculated by dividing the point-to-point glucose
consumption or lactate production differences by the point-to-point
differences in time; ##EQU18## wherein: q.sub.i is the specific
glucose consumption or lactate production rate (.mu.mole/cell/hr)
at time t.sub.i ; and
t.sub.i, Q.sub.i and X.sub.i are defined above;
determining the percentage of colony-forming cells (% CFC) in the
reference culture at each of the selected times;
using the % CFC and q.sub.lac or q.sub.gluc to determine
.alpha..sub.lac and .beta..sub.lac or .alpha..sub.gluc and
.beta..sub.gluc using equation (6) or (9) set forth below:
##EQU19## wherein: q.sub.lac is the specific lactate production
rate;
.alpha..sub.lac is the q.sub.lac value for a CFC; and
.beta..sub.lac is the q.sub.lac value for a non-CFC; ##EQU20##
wherein: q.sub.gluc is the specific glucose consumption rate;
.alpha..sub.gluc is the q.sub.gluc value for a CFC; and
.beta..sub.gluc is the q.sub.gluc value for a non-CFC;
culturing an experimental culture of hcmatopoietic cells under
essentiallythe same conditions used to culture the reference
culture;
measuring the total glucose consumption or lactate production per
unit volume of the experimental culture at a plurality of selected
times;
measuring the density of nucleated cells (X.sub.i) present in the
experimental culture at the selected times;
using the measured X.sub.i and the measured glucose consumption or
lactate production to calculate q.sub.gluc or q.sub.lac for the
experimental culture using equations (1) and (2); and
using the calculated q.sub.lac or q.sub.gluc for the experimental
culture and the values of .alpha..sub.lac and .beta..sub.lac or
.alpha..sub.gluc and .beta..sub.gluc for the reference culture to
calculate the % CFC at the selected times for the experimental
culture using equation (6) or (9).
17. The method of claim 16 wherein lactate production is
measured.
18. A method of determiining colony-forming cell content of a
hematopoietic cell culture, the method comprising:
culturing a reference culture of hematopoietic cells under selected
conditions, the selected conditions comprising using a feeding
protocol that limits the production of post-progenitor cells that
have high glucose consumption and lactate production;
measuring the total glucose consumption or lactate production per
unit volume of the reference culture at a plurality of selected
times;
measuring the density of nucleated cells (X.sub.i) present in the
reference culture at each of the selected times;
using the measured X.sub.i and the measured glucose consumption or
lactate production to calculate q.sub.gluc or q.sub.lac at each of
the selected times using equations (1) and (2) set forth below:
##EQU21## wherein: Q.sub.i is the total glucose consumed or lactate
produced (.mu.mole/ml/hr) at time t.sub.i ;
t.sub.i is any time point;
t.sub.i+1 is a time point after time point t.sub.i ;
t.sub.i-1 is a time point before time point t.sub.i ;
slope.sub.b is the first order backward slope from time t.sub.i-1
to time t.sub.i of the total glucose consumption or lactate
production curve calculated by dividing the point-to-point glucose
consumption or lactate production differences by the point-to-point
differences in time; and
slope.sub.f is the first order forward slope from time t.sub.i to
time t.sub.i+1 of the total glucose consumption or lactate
production curve calculated by dividing the point-to-point glucose
consumption or lactate production differences by the point-to-point
differences in time; ##EQU22## wherein: q.sub.i is the specific
glucose consumption or lactate production rate (.mu.mole/cell/hr)
at time t.sub.i ; and
t.sub.i, Q.sub.i and X.sub.i are defined above;
determining the percentage of colony-forming cells (% CFC) in the
reference culture at each of the selected times;
using the % CFC and q.sub.lac or q.sub.gluc to determine
.alpha..sub.lac and .beta..sub.lac or .alpha..sub.gluc and
.beta..sub.gluc using equation (6) or (9) set forth below:
##EQU23## wherein: q.sub.lac is the specific lactate production
rate;
.alpha..sub.lac is the q.sub.lac value for a CFC; and
.beta..sub.lac is the q.sub.lac value for a non-CFC; ##EQU24##
wherein: q.sub.gluc is the specific glucose consumption rate;
.alpha..sub.gluc is the q.sub.gluc value for a CFC; and
.beta..sub.gluc is the q.sub.gluc value for a non-CFC;
culturing an experimental culture of hematopoietic cells under
essentially the same conditions used to culture the reference
culture;
measuring the total glucose consumption or lactate production per
unit volume of the experimental culture at a plurality of selected
times;
using the measured glucose consumption or lactate production to
calculate Q.sub.gluc or Q.sub.lac at the selected times using
equation (1);
measuring the concentration of nucleated cells (X.sub.i) present in
the experimental culture at the selected times; and
using the measured X.sub.i and the calculated Q.sub.lac or
Q.sub.gluc for the experimental culture and the values of
.alpha..sub.lac and .beta..sub.lac or .alpha..sub.gluc and
.beta..sub.gluc for the reference culture to calculate the
concentration of total CFC at the selected times for the
experimental culture using equation (4) or (7) set forth below:
wherein:
Q.sub.lac is the total lactate production;
is the concentration of total CFC in the culture; and
.alpha..sub.lac, .beta..sub.lac and X.sub.i are defined above;
wherein:
Q.sub.gluc is total glucose consumption; and
.alpha..sub.gluc, .beta..sub.gluc, X.sub.i and are defined
above.
19. A method of determining colony-forming cell content of a
hematopoietic cell culture, the method comprising:
culturing a reference culture of hematopoietic cells under selected
conditions, the selected conditions comprising employing a selected
feeding protocol;
measuring the volumetric oxygen uptake rate (Q.sub.02)
(.mu.mole/ml/hr) of the reference culture at a plurality of
selected times;
measuring the density of nucleated cells (X.sub.i) present in the
reference culture at each of the selected times;
using the measured X.sub.i and Q.sub.02 to calculate q.sub.02 at
each of the selected times using equation (17) set forth below:
wherein:
q.sub.02 is the specific oxygen consumption rate
(.mu.mole/cell/hr); and
Q.sub.02 and X.sub.i are defined above;
determining the percentage of colony-forming cells (% CFC) in the
reference culture at each of the selected times;
using the % CFC and q.sub.02 to determine .alpha..sub.02 and
P.sub.02 using equation (20) set forth below: ##EQU25## wherein:
.alpha..sub.02 is the q.sub.02 value for a CFC;
.beta..sub.02 is the q.sub.02 value for a non-CFC; and
q.sub.02 is defined above;
culturing an experimental culture of hematopoietic cells under
essentially the same conditions used to culture the reference
culture;
measuring the Q.sub.02 of the experimental culture at one or more
selected time(s);
measuring the density of nucleated cells (X.sub.i) present in the
experimental culture at the selected time(s);
using the measured X.sub.i and Q.sub.02 to calculate q.sub.02 for
the experimental culture using equation (17); and
using the calculated q.sub.02 for the experimental culture and the
values of .alpha..sub.02 and .beta..sub.02 for the reference
culture to calculate the % CFC at the selected time(s) for the
experimental culture using equation (20).
20. A method of determining colony-forming cell content of a
hematopoietic cell culture, the method comprising:
culturing a reference culture of hematopoietic cells under selected
conditions, the selected conditions comprising employing a selected
feeding protocol;
measuring the volumetric oxygen uptake (Q.sub.02) (.mu.mole/ml/hr)
of the reference culture at a plurality of selected times;
measuring the density of nucleated cells (X.sub.i) present in the
reference culture at each of the selected times;
using the measured X.sub.i and Q.sub.02 to calculate q.sub.02 at
each of the selected times using equation (17) set forth below:
wherein:
q.sub.02 is the specific oxygen consumption rate
(.mu.mole/cell/hr); and
Q.sub.02 and X.sub.i are defined above;
determining the percentage of colony-forming cells (% CFC) in the
reference culture at each of the selected times;
using the % CFC and q.sub.02 to determine .alpha..sub.02 and
.beta..sub.02 using equation (20) set forth below: ##EQU26##
wherein: .alpha..sub.02 is the q.sub.02 value for a CFC;
.beta..sub.02 is the q.sub.02 value for a non-CFC; and
q.sub.02 is defined above;
culturing an experimental culture of hematopoietic cells under
essentially the same conditions used to culture the reference
culture;
measuring the Q.sub.02 of the experimental culture at one or more
selected time(s);
measuring the concentration of nucleated cells (X.sub.i) present in
the experimental culture at the selected time(s); and
using the measured X.sub.i and Q.sub.02 for the experimental
culture and the values of .alpha..sub.02 and .beta..sub.02 for the
reference culture to calculate the concentration of total CFC at
the selected time(s) for the experimental culture using equation
(21) set forth below:
wherein:
.alpha..sub.02 is the q.sub.02 value for a CFC;
.beta..sub.02 is the q.sub.02 value for a non-CFC;
is the concentration of total CFC in the culture; and
q.sub.02, Q.sub.02 and X.sub.i are defined above.
Description
FIELD OF THE INVENTION
The invention relates to methods of determining the content of
progenitor cells (also referred to as colony forming cells) in a
hematopoietic cell culture. The method of the invention utilizes
measurements of one or more metabolic parameters of the culture,
such as glucose consumption, oxygen consumption, and lactate
production, to determine the progenitor cell content.
BACKGROUND
The culture of hematopoietic cells for transplantation therapies is
a rapidly growing area of biotechnology and experimental
hematology. As evidenced by recent clinical trials (Brugger et al.,
New Engl. J. Med., 333, 283-287 (1995); Williams et al., Blood, 87,
1687-1691 (1996); Bertolini et al., Blood, 89, 2679-2688 (1997)),
ex vivo expanded hematopoietic cells offer great promise for the
reconstitution of in vivo hematopoiesis in patients who have
undergone chemotherapy. Other potential applications for ex vivo
expansion include production of cycling stem and progenitor cells
for gene therapy, expansion of dendritic cells for immunotherapy,
and production of red blood cells and platelets for transfusions
(McAdams et al., Trends Biotechnol., 14, 388-396 (1996)). Thus, it
is likely that the demand for ex vivo expanded hematopoietic cells
will increase dramatically.
Hematopoietic cultures are among the most challenging culture
systems. The heterogeneous cell population contained in a
hematopoietic culture is always changing as a result of the
delicate balance between proliferation of certain cell types, their
differentiation into other cell types, and the death of various
cell populations. The lifespan of cells in culture is likely to
depend on cytokine stimulation, as well as on a number of
physicochemical parameters, such as pH, dissolved oxygen, and
nutrient and metabolite concentrations (McAdams et al., Trends
Biotechnol., 14, 341-349 (1996)).
Current enumeration techniques for hematopoietic cultures do not
provide real-time analysis of the changing populations. Complete
evaluation of the performance of hematopoietic cultures requires
the use of assays with long durations, such as the two-week
methylcellulose assay to detect progenitor or colony-forming cells
(CFC), including colony-forming units-granulocyte/monocyte (CFU-GM)
and burst-forming units-erythroid (BFU-E), and the seven-week assay
for the very primitive long-term culture-initiating cells (LTC-IC).
In this regard, the cell requirements for successful engraftment
are often expressed in terms of the number of CFC transplanted
(e.g., 2.times.10.sup.5 CFU-GM per kg body weight; Bender et al.,
J. Hematotherapy, 1, 329-341 (1992)). In contrast to the long assay
times, the time period available to determine when to harvest
ex-vivo cultures for transplantation therapies is most likely on
the order of hours. Currently, only flow cytometry offers this
speed of analysis. Flow cytometry can be utilized to quantify cells
bearing antigens such as CD34 (primitive progenitors), CD15 and
CD11b (granulocyte and monocyte post-progenitors), and gly A
(maturing erythrocytes). Even so, sample preparation and
measurement, along with data analysis, requires 2-3 hours.
Furthermore, when cells bearing the antigen of interest are present
at a low concentration, as is often the case for CFC, accurate
quantitation may be difficult. Also, it should be noted that,
although most CFC present in hematopoietic cell sources (e.g., bone
marrow or umbilical cord blood) express the CD34 antigen, CD34
expression by cultured cells is often lost before the CFC content
of a culture is depleted. Because of the difficulty in determining
when CFU-GM, BFU-E, or other cell populations of interest reach a
maximum level, culture endpoints have generally been chosen based
on a retrospective analysis of the culture duration that usually
yields an acceptable product. While the use of retrospective
analysis may be adequate, it is far from optimal due to the
heterogeneity in the kinetics of cell expansion (e.g., initial
quiescent phase and the time at which various cell populations
reach a maximum) for different hematopoietic cell source
samples.
Nutrient consumption and by-product accumulation rates are
parameters that can be readily measured in real-time. These rates
are frequently employed for the control of more traditional cell
cultures (for vaccine and protein production), but have been
largely overlooked in the evaluation of hematopoietic cultures.
Normal and leukemic human blood cells depend heavily upon
glycolysis as their source of energy (Beck, J. Biol. Chem., 232,
251-270 (1958); Beck and Valentine, Cancer Res., 12, 818-822
(1952); Beck and Valentine, Cancer Res., 12, 823-828 (1952)), and
the rates of glucose consumption and lactate production can be
altered by external stimuli, such as growth factors. Growth
factor-stimulated increases in glucose utilization have been
demonstrated in cultures of murine macrophages (Hamilton et al.,
Biochem. Biophys. Res. Commun., 138, 445-454 (1986)) and
multipotential hematopoietic cell lines (Whetton et al., EMBO J.,
3, 409-413 (1984); Whetton, et al., J. Cell Sci., 84, 93-104
(1986)). Human lymphocytes stimulated to undergo blastogenesis by
incubation with phytohemagglutinin (PHA) exhibit increased glucose
utilization and lactate production and increased levels of
glycolytic pathway enzymes (Hedeskov, Biochem. J., 110, 373-380
(1968); Rogers et al., Ann. Hum. Genet., 43, 213-226 (1980); Kester
et al., Arch. Biochem. Biophys., 183, 700-709 (1977)). The findings
discussed above for stimulated hematopoietic cells are consistent
with those for rapidly dividing cells in general, which are known
to exhibit rates of glucose consumption and lactate production that
are elevated over those of more slowly growing cells (Hume and
Weidemann, J. Natl. Cancer
Inst., 62, 3-8 (1979); Newsholme, et al., Biosci. Rep., 5, 393-400
(1985); Lanks and Li, J. Cell. Physiol., 135, 151-155 (1988)). Data
regarding oxygen consumption rates in human hematopoietic cultures
are scarce, and the published reports have not fully examined the
effects of various cell populations on oxygen metabolism. Peng and
Palsson (Annals Of Biomedical Engineering, 24, 373-381 (1996))
examined oxygen uptake by human bone marrow cells in modified
six-well culture plates. They found that the specific oxygen uptake
rate (moles per cell per hour) increased steadily during the first
10 days in culture and then remained steady or increased slightly
from days 10-14. Bird et al. (Cancer, 1009-1014 (1951)) measured
oxygen uptake by normal human granulocytes.
SUMMARY OF THE INVENTION
In view of the foregoing, it was expected, in a hematopoietic
culture containing cells with varying proliferative potential and
rates of growth, that metabolic parameters of the culture, such as
glucose consumption, oxygen consumption and lactate production,
would vary directly with total cell density as proliferation and
differentiation occurred. Instead, it has surprisingly been
discovered that the metabolic parameters of a hematopoietic cell
culture correlate with the progenitor cell (colony forming cell)
content of such a culture.
Accordingly, the invention provides a method of determining the
content of progenitor cells in a hematopoietic cell culture. The
method comprises culturing a reference culture of hematopoietic
cells under selected conditions. At least one metabolic parameter
of the reference culture is measured at a plurality of selected
times. Also, the content of progenitor cells in the reference
culture is determined at each of the selected times. Then, an
experimental culture of hematopoietic cells is cultured under
essentially the same conditions used to culture the reference
culture, and the same metabolic parameter(s) as was(were) measured
for the reference culture is(are) measured for the experimental
culture at a selected time. Finally, the metabolic parameter(s)
measured for the experimental culture is(are) compared with that
(those) measured for the reference culture to determine the content
of the progenitor cells in the experimental culture at the selected
time.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1: Time profile for specific glucose consumption rate
(q.sub.gluc) (....box-solid....), specific lactate production rate
(q.sub.lac) (....circle-solid....), and the % of cells that are
CFU-GM (-.largecircle.-), BFU-E (-.DELTA.-), and total CFC
(CFU-GM+BFU-E) (-----) in a spinner flask. The cord blood (CB)
mononuclear cell (MNC) culture was carried out in XVIVO-20 medium
with interleukin-3 (IL-3), interleukin-6 (IL-6), stem cell factor
(SCF), granulocyte-colony stimulating factor (G-CSF),
granulocyte-macrophage colony stimulating factor (GM-CSF), and
erythropoietin (Epo) at an inoculum density (ID) of
3.8.times.10.sup.5 cells/ml.
FIG. 2: Time profile for q.sub.lac and the % of cells that are
CFU-GM, BFU-E and total CFC (CFU-GM+BFU-E) in a T-flask. The sample
and culture conditions used and the symbols are the same as in FIG.
1.
FIGS. 3A-C: Time profiles for q.sub.lac and the % of cells that are
CFU-GM, BFU-E and total CFC (CFU-GM+BFU-E) in spinner flasks for a
single peripheral blood (PB) MNC sample cultured at an ID of
160,000 cells/ml (FIG. 3A), 750,000 cells/ml (FIG. 3B), and
1,260,000 cells/ml (FIG. 3C). All cultures were carried out in
human long-term medium (HTLM) with IL-3, IL-6, SCF, G-CSF, GM-CSF,
and Epo. The symbols are the same as in FIG. 1.
FIGS. 4A-B: Time profiles for q.sub.lac and the % of cells that are
CFU-GM, BFU-E and total CFC (CFU-GM+BFU-E) in spinner flasks for a
single PB MNC sample cultured at an ID of 1,260,000 cells/ml in
XVIVO-20 with IL-3, IL-6, SCF, G-CSF and either GM-CSF and Epo
(FIG. 4A), or Flt3 ligand (Flt3-l) (FIG. 4B). The symbols are the
same as in FIG. 1.
FIGS. 5A-B: Time profiles for q.sub.lac and the % of cells that are
CFU-GM, BFU-E and total CFC (CFU-GM+BFU-E) in spinner flasks for a
single PB MNC sample cultured at an ID of 800,000 cells/ml in
XVIVO-20 with IL-3, IL-6, SCF and either Epo (FIG. 5A), or G-CSF
(FIG. 5B). The symbols are the same as in FIG. 1.
FIGS. 6A-C: Time profiles for q.sub.lac and the % of cells that are
CFU-GM, BFU-E and total CFC (CFU-GM+BFU-E) in T-flasks for a single
PB CD34.sup.+ cell sample cultured at an ID of 33,000 cells/ml
(FIG. 6A), 82,000 cells/ml (FIG. 6B), and 125,000 cells/ml (FIG.
6C). The cultures were conducted in HLTM with IL-3, IL-6, SCF,
G-CSF, GM-CSF, and Epo. The symbols are the same as in FIG. 1.
FIG. 7: Graph of q.sub.lac vs. % CFC for a CB MNC culture carried
out in a spinner flask at an ID of 250,000 cells/ml in HLTM plus
IL-3, IL-6, SCF, G-CSF, GM-CSF, and Epo. The correlation
coefficient for the line is 0.99. The error bars indicate the
mean.+-.one standard deviation. Standard deviations were determined
by propagation of the errors in cell density (5%), lactate
concentration (2% or 5%), and CFC ( ##EQU1## where N is the total
number of colonies counted) through the calculations for q.sub.lac
and % CFC.
FIGS. 8A-C: Graph of q.sub.lac vs. % CFC in spinner flasks for a
single PB MNC sample cultured at an ID of 160,000 cells/ml (FIG.
8A), 750,000 cells/ml (FIG. 8B), and 1,260,000 cells/ml (FIG. 8C).
The cultures were carried out in HLTM plus IL-3, IL-6, SCF, G-CSF,
GM-CSF, and Epo. The data are the same as those shown in
time-course format in FIG. 3. The correlation coefficients for the
lines are: 0.96 (FIG. 8A), 0.63 (FIG. 8B), and 0.74 (FIG. 8C). The
error bars indicate the mean.+-.one standard deviation and were
determined as described for FIG. 7. The value of .alpha. was
4.66x10.sup.-7, and the value of .beta. was 3.41x10.sup.-8.
FIGS. 9A-C: Graph of q.sub.lac vs. % CFC in T-flasks for a single
PB CD34.sup.+ cell sample cultured at an ID of 33,000 cells/ml
(FIG. 9A), 82,000 cells/ml (FIG. 9B), and 125,000 cells/ml (FIG.
9C). The cultures were carried out in HLTM plus IL-3, IL-6, SCF,
G-CSF, GM-CSF, and Epo. The data are the same as those shown in
time-course format in FIG. 6, except that the first time point has
been deleted from each culture. The correlation coefficients for
the lines are 0.94 (FIG. 9A), 0.79 (FIG. 9B), and 0.95 (FIG. 9C).
The error bars indicate the mean.+-.one standard deviation and were
determined as described for FIG. 7.
FIG. 10: Graph of q.sub.lac vs. % CFC for three PB MNC cultures
carried out in spinner flasks with HLTM plus IL-3, IL-6, SCF,
G-CSF, GM-CSF, and Epo at an ID of 1.2.times.10.sup.6 cells/ml. The
day zero CD34.sup.+ cell content of the samples were 6.6%
(.circle-solid.), 6.3% (.quadrature.), and 5.8% (.DELTA.). The
correlation coefficient for the line is 0.88. The error bars
indicate the mean.+-.one standard deviation and were determined as
described in FIG. 7.
FIGS. 11A-B: Time profiles for q.sub.gluc and the % of cells that
are CFU-GM, BFU-E and total CFC for a PB MNC sample cultured at an
ID of 700,000 cells/ml (FIG. 11A) and 450,000 cells/ml (FIG. 11B).
All cultures were carried out with IL-3, IL-6, SCF, G-CSF, GM-CSF,
and Epo in either HTIM medium (FIG. 11A) or XVIVO-20 medium (FIG.
11B). In FIGS. 11A-B, .largecircle.=%GM, .increment.=%,
.circle-solid.=% CFC, and the dashed line is q.sub.gluc.
FIGS. 12A-B: Graph of q.sub.gluc vs. % CFC for the same cultures as
FIGS. 11A-B. ID of 700,000 cells/ml in HTLM medium (FIG. 11A), and
ID of 450,000 cells/ml in XVIVO-20 medium (FIG. 11B). In FIG. 12A,,
y=2.61E-06x+3.86E-08 and R.sup.2 =9.28E-01; and in FIG. 12B,
y=7.69E-07x+2.11E-08 and R.sup.2 =8.86E-01.
FIGS. 13A-B: Time profiles for the expansion of total cells (FIG.
13A) and CFU-GM (FIG. 13B) for a single CB MNC sample (Experiment
1) cultured in a T-flask (-.tangle-solidup.-) and a bioreactor
(-.largecircle.-) at an ID of 1.37.times.10.sup.6 cells/ml. The
final culture density was 6.2.times.10.sup.6 cells/ml in the
bioreactor and 8.8.times.10.sup.6 cells/ml in the T-flask. The
cultures were conducted in HLTM with IL-3, IL-6, SCF, G-CSF,
GM-CSF, and Epo and were fed using a cell-retention feeding
protocol (FP1), except that the bioreactor was diluted from
5.2.times.10.sup.6 to 3.2.times.10.sup.6 cells/ml at 144 hours.
FIG. 14: Time profile of the cell density for a single CB MNC
sample (Experiment 3) cultured in a T-flask (-.tangle-solidup.-)
and in a bioreactor (-.largecircle.-) at an ID of
1.3.times.10.sup.6 cells/ml. The cultures were conducted in HLTM
with IL-3, IL-6, SCF, G-CSF, GM-CSF, and Epo and were fed using a
dilution feeding protocol (FP2).
FIGS. 15A-C: Time profiles for total cell expansion in a T-flask
(-.tangle-solidup.-) and a bioreactor (-.largecircle.-) using FP2
for Experiment 2 (CB MNC, ID=1.72.times.10.sup.6 cells/ml) (FIG.
15A), Experiment 3 (CB MNC, ID=1.3.times.10.sup.6 cells/ml) (FIG.
15B), and Experiment 4 (PB MNC, ID=2.24.times.10.sup.5 cells/ml)
(FIG. 15C).
FIGS. 16A-C: Time profiles for CFU-GM expansion in a T-flask
(-.tangle-solidup.-) and a bioreactor (-.largecircle.-) with FP2
for Experiment 2 (FIG. 16A), Experiment 3 (FIG. 16B), and
Experiment 4 (FIG. 16C).
FIGS. 17A-D: Time profiles for % CFC (-.largecircle.-), q.sub.lac
(-.DELTA.-), and q.sub.02 (-.tangle-soliddn.-) for Experiment 1
(FIG. 17A), Experiment 2 (FIG. 17B), Experiment 3 (FIG. 17C), and
Experiment 4 (FIG. 17D).
FIGS. 18A-D: Time profiles for % CFC (-.largecircle.-),
Y.sub.lac,ox (-.gradient.-), and % CD11b.sup.+ and/or CD15.sup.+
cells (-.circle-solid.-) for Experiment 1 (FIG. 18A), Experiment 2
(FIG. 18B), Experiment 3 (FIG. 18C), and Experiment 4 (FIG.
18D).
FIGS. 19A-D: Graphs of q.sub.02 versus % CFC in the bioreactor for
Experiment 1 (FIG. 19A) (y=1.73.times.10.sup.-8
.times.+3.98.times.10.sup.-8 ; r.sup.2 =0.87), Experiment 2 (FIG.
19B) (y=5.37.times.10.sup.-7 .times.+4.5.times.10.sup.-8 ; r.sup.2
=0.66), Experiment 3 (FIG. 19C) (y=5.5.times.10.sup.-7
.times.+5.96.times.10.sup.-8 ; r.sup.2 =0.8) and Experiment 4 (FIG.
19D) (y=6.12.times.10.sup.-7 .times.+3.7.times.10.sup.-8 ; r.sup.2
=0.86).
DETAILED DESCRIPTION OF THE PRESENTLY PREFERRED EMBODIMENTS
A very small population of totipotent stem cells proliferate and
differentiate to produce all blood-cell lineages in the body. Less
primitive pluripotent stem cells may differentiate only into a
subset of lineages. The intermediate-stage cells that are committed
to specific lineages, but are still capable of significant
proliferation, are known as progenitor cells. Colony-forming cell
assays are the most common type of in vitro assays for progenitor
cells, and progenitor cells are also referred to as colony-forming
cells (CFC). The final functional cells of each lineage are termed
mature blood cells. The mature blood cells include lymphocytes,
erythrocytes, neutrophils, macrophages, dendritic cells, and
platelets. The mature cells and all of the precursors of mature
blood cells (including all of the stem and progenitor cell
populations) are termed hematopoietic cells.
In the method of the invention, a reference culture of
hematopoietic cells is cultured under selected conditions. Methods
of culturing hematopoietic cells and the conditions that affect
such cultures are well known. See McAdams et al., Trends
Biotechnol., 14, 341-349 (1996) for a review of these methods and
conditions. Such conditions include the source of the cells,
inoculum density, identity of cytokines used, and physicochemical
conditions. The selected conditions are a set of conditions
selected for a particular culture which are effective, preferably
optimal, for the growth of the hematopoietic cell culture as a
whole, or of a particular type of hematopoietic cell. Such
effective and optimal conditions are known in the art or can be
determined empirically. Making such determinations is within the
skill in the art.
There are three main sources of hematopoietic cells for use in the
methods of the invention. These are bone marrow (BM), umbilical
cord blood (CB) and peripheral blood (PB). While BM has been the
traditional source of hematopoietic cells for transplantation
therapies, other blood cell sources are becoming more popular.
Mobilized PB progenitor cell transplants have proved effective, and
it is likely that mobilized PB mononuclear cells (MNC) will replace
BM MNC as the preferred source of hematopoietic cells for
transplantation (Korbling and Champlin, Stem Cells, 14, 185-195
(1996)). CB is both readily available and easily collected. CB stem
cells are thought to be more immature than those found in adults.
This attribute makes CB stem cells a potential target for the
correction of genetic blood diseases (Clapp and Williams, Stem
Cells, 13, 613-621 (1995)).
Methods of collecting BM, CB, and PB are well known in the art.
See, e.g., McAdams et al., Trends Biotechnol., 14, 341-349 (1996).
Briefly, BM is collected under general anesthesia by multiple
needle aspirations to the sternum and/or pelvis. CB is obtained
non-invasively from the umbilical cord of newborn infants. BM and
CB are often processed by centrifugation over a Ficoll-Histopaque
density gradient to deplete the sample of erythrocytes. BM and CB
can be cultured without this processing, but high red blood cell
content can make visual observation of culture growth quite
difficult. PB cells are generally obtained after stem cell
mobilization, which is achieved by the administration of one or
more of several chemotherapeutic drugs and/or growth factors (often
granulocyte colony stimulating factor (G-CSF)) to the patient. By
an as yet unknown mechanism, these drugs and factors cause a large
number of stem and progenitor cells to proliferate and/or exit the
BM and enter the peripheral circulation. Nucleated cells are then
collected from the patients using a blood-processing machine; this
collection process is known as peripheral blood apheresis.
Purified CD34.sup.+ cells can be used in the cultures of the
invention. CD34 is a surface glycoprotein of unknown function that
is found on approximately 1% (0.1-10%) of collected hematopoietic
mononuclear cells (MNCs). It is present on all of the most
primitive cells, from the quiescent stem cells to the highly
proliferative progenitor cells. As nearly all of the proliferative
potential initially present in hematopoietic cell cultures is
represented by the CD34.sup.+ cells, a number of methods have been
developed for their selection. All of these methods rely upon the
use of an antibody which recognizes the CD34 antigen and subsequent
recovery of the cell-antibody complex. For instance, the cells may
be treated with a hapten-conjugated anti-CD34 antibody, and then
collected by attachment to anti-hapten antibodies coupled to
adsorption columns or magnetic beads. These methods will provide
cell populations enriched in CD34.sup.+ cells, but the degree of
purity will vary depending on the method used. For applications
which require extremely high purity (>95%) CD34.sup.+ cells,
fluorescence-activated cell sorting (FACS) is recommended. See de
Wynter et al., Stem Cells, 13, 524-532 (1995) for a review of
CD34.sup.+ isolation techniques. See also Papadimitriou et al., J.
Hematotherapy, 4, 539-544 (1995); Winslow et al., Bone Marrow
Transplant, 14, 265-271 (1994). When compared with MNCs, cultures
initiated with CD34.sup.+ cells have much greater expansion
potential. However, CD34.sup.+ selection is often expensive and
often results in significant cell loss. CD34.sup.+ cell populations
also lack accessory cells, such as macrophages, that may provide
cytokines and other stimulatory factors in MNC cultures.
The expansion potential of different sources of hematopoietic
cells, and even of different samples of the same type of cells, is
usually subject to large variations of up to several orders of
magnitude. CB has greater expansion potential than the other
sources, while PB has the greatest variation owing to the wide
variety of mobilization regimens and disease states of the patients
from whom samples are taken. Also, as noted above, CD34.sup.+ cells
have greater expansion potential than MNCs. The kinetics of cell
expansion also differ greatly between different samples of the same
types of cells.
The invention will allow those skilled in the art to estimate the
progenitor cell (CFC) content in cultures with differing expansion
kinetics and potential. However, the cells in the reference and
experimental cultures must be from the same source (e.g., CB), must
be from similar samples (e.g., samples from similar donors
(age-matched, same
disease states, same treatments)) or samples which have similar
CD34.sup.+ contents, and must be processed in the same way once
removed from the donor. For instance, if PB is the source of the
cells, the reference culture must employ PB cells from a patient
with the same disease state as the source of the cells for the
experimental culture and the same mobilization regimen must be
used, or the reference and experimental cultures must be initiated
from samples having similar contents of CD34.sup.+ cells. As
another example, if purified CD34.sup.+ cells are used in the
experimental culture, purified CD34.sup.+ cells must also be used
in the reference culture. Also, if frozen samples of cells are used
in the experimental culture, then frozen samples must be used in
the reference culture.
The inoculum density refers to the quantity of MNCs or CD34.sup.+
cells per unit volume in the cell population used to initiate the
cultures. Inoculum densities based on CD34.sup.+ cells can be, and
are preferably, used even when the cells are not purified
CD34.sup.+ cells. The use of CD34.sup.+ inoculum densities is
preferable because CD34.sup.+ cells are much more proliferative
than are more mature C34.sup.- cells. It is, therefore, expected
that standardization with respect to CD34.sup.+ cell content will
give greater reproducibility and more reliable results.
Methods of determining the number of MNCs and CD34.sup.+ cells
present in a sample are well known in the art. Preferably, the MNCs
are counted using a Coulter Counter or similar apparatus (see
Example 1). Preferably, the percentage CD34.sup.+ cells is
determined by flow cytometry.
For static cultures, the inoculum density of MNCs should not be
below 5.times.10.sup.4 cells/ml or above 5.times.10.sup.5 cells/ml.
For CD34.sup.+ cell cultures, the inoculum density should not be
below 2.times.10.sup.4 cells/ml and should not exceed
5.times.10.sup.4 cells/ml. Higher inoculum densities will deplete
key nutrients too quickly, necessitating frequent feeding of the
cultures, and lower densities will not provide reproducible cell
expansion. In general, within the limits given above, lower density
cultures will exhibit a greater expansion of total cells and
progenitor cells than higher density cultures. However, if a large
number of cells are required, higher density cultures are
recommended since higher densities produce greater total cells and
CFCs.
For stirred cultures carried out in spinner flasks utilizing a
serum-containing medium, the inoculum density for MNCs should be no
lower than 2.times.10.sup.5 cells/ml. For serum-free cultures, the
inoculum density should be no lower than 3.times.10.sup.5 cells/ml.
For CD34.sup.+ cultures, the inoculum density should be no lower
than 5.times.10.sup.4 cells/ml for either type of medium. MNC
cultures as high as 1.5.times.10.sup.6 cells/ml have been initiated
with good results. Inoculum densities for CD34.sup.+ cultures
should not exceed 1.2.times.10.sup.5 cells/ml. A good intermediate
inoculum density for both serum-containing and serum-free cultures
is 5.times.10.sup.5 cells/ml for MNC cultures and
7.5.times.10.sup.4 cells/ml for CD34.sup.+ cultures.
For stirred cultures carried out in a bioreactor, the optimal
inoculum density ranges have not yet been established. However,
they are expected to be similar to those for spinner flask
cultures. For example, MNC cultures in bioreactors in
serum-containing medium have been initiated with good results using
inoculum densities of 0.2-1.7.times.10.sup.6 cells/ml. As used
herein, "bioreactor" means any culture vessel which provides a
full-instrumented, well-controlled, closed and reproducible culture
environment.
Cytokines must be included in hematopoietic cell cultures to obtain
proliferation and differentiation of hematopoietic cells. Suitable
cytokines, their properties, and guidelines for their use in
hematopoietic cultures are known in the art. See Sui et al., Proc.
Natl. Acad. Sci. (USA), 92, 2859-2863 (1995); Farese et al., Blood,
87, 581-591 (1996); Gore et al., Exp. Hematol., 23, 413-421 (1995);
Mayani et al., Blood, 81, 3252-3258 (1993); Sonoda et al., Proc.
Natl. Acad. Sci. (USA), 85, 4360-4364 (1988); Tanaka et al., Blood,
86, 73-79 (1995); Massague and Pandiella, Ann. Rev. Biochem., 62,
515-541 (1993); Nathan and Sporn, J. Cell. Biol., 113, 981-986
(1991); Nicola, Ann. Rev. Biochem., 58, 45-77 (1989). Also, it is
expected that new cytokines will be discovered or developed which
can be used in the methods of the invention. Effective and optimal
concentrations of cytokines for use in hematopoietic cultures are
known in the art or can be determined empirically, and making such
determinations is within the skill in the art. Additionally, by
appropriate choice of the cytokines used in the culture, broad
expansion across multiple hematopoietic lineages or expansion of a
specific lineage of cells can be obtained. The chosen cytokine(s)
should produce an increase in total cells whether multiple lineages
or a single lineage is expanded.
Cytokines can be classified into three groups:
i. a group acting on primitive hematopoietic cells (e.g. stem cell
factor, Flt3 ligand);
ii. a group acting on a wide array of progenitor cells (e.g.
interleukin-3, interleukin-6, granulocyte-macrophage colony
stimulating factor, PIXY321); and
iii. a group acting on more mature, lineage-restricted cells (e.g.
granulocyte colony stimulating factor, macrophage colony
stimulating factor, erythropoietin, thrombopoietin).
Depending on the desired culture product, a combination of these
cytokines is typically utilized. Most expansion protocols use a
combination with at least one cytokine from each of groups i and ii
and perhaps one cytokine from group iii. For instance, a protocol
for granulocyte production might utilize a combination of stem cell
factor, interleukin-3, interleukin-6, and granulocyte colony
stimulating factor.
The above list of cytokines is by no means exhaustive. There are
several additional cytokines that can be used. See, e.g., McAdams
et al., Trends Biotechnol., 14, 341-359 (1996); McKenna et al.,
Blood, 86, 3413-3420 (1995); and Debili et al., Blood, 86,
2516-2525 (1995). Also, "designer" cytokines have been developed
which combine the active regions of two cytokines or mimic the
binding domain of the cytokine receptor ligand and which have
enhanced or novel activities. See, e.g., McAdams et al., Trends
Biotechnol., 14, 341-359 (1996). As noted above, it is expected
that new cytokines will be discovered or developed which can be
used in the methods of the invention.
The physicochemical conditions that affect hematopoietic cell
cultures are well known and include the culture medium, pH,
incubation conditions (e.g., atmosphere and temperature), type of
culture vessel, feeding schedules, biocompatibility of tissue
culture materials, culture system (e.g., stirred versus static),
etc. For a review of these physicochemical conditions and their
effects on hematopoietic cell cultures, see McAdams, et al., Trends
Biotechnol., 14, 341-349 (1996). Some of these conditions will be
discussed briefly.
Either serum-containing or serum-free medium can be used in
hematopoietic cultures. Serum-containing medium generally gives
higher progenitor cell and total cell expansion. However, if a more
defined medium is desired (e.g., for clinical applications),
acceptable expansion can be obtained using serum-free medium. Also,
serum-containing medium favors the expansion and maturation of the
granulocyte and macrophage lineages, while serum-free medium
promotes greater expansion of the erythroid and megakaryocyte
lineages. Equilibrating the medium with the incubation atmosphere
and temperature is recommended prior to adding the cells to the
medium.
The biocompatibility of materials is an important issue in
hematopoeitic cultures. Tissue culture treated polystyrene,
commonly utilized in the construction of well plates and T-flasks,
is biocompatible with hematopoietic cells. However, other materials
commonly used for the construction of culture devices for animal
cells may not be compatible with hematopoietic cells. Silicone,
glass and polycarbonate are a few of the materials which adversely
affect hematopoietic culture performance, as identified in a recent
publication. LaIuppa, J. Biomed. Mat. Res., 36, 347-359 (1997). See
also, McAdams et al., Trends Biotechnol., 14, 341-349 (1996). When
designing culture systems, the performance of hematopoietic cells
on a chosen material should be evaluated before use of the material
in cultures. Material compatibility is especially important if the
culture is carried out in serum-free medium; serum can partially
protect hematopoietic cells from the negative effects of some
materials.
Oxygen tension plays a significant role in hematopoietic culture
performance. The oxygen tension of the gas in the headspace of the
culture is typically 5-20%. Studies suggest that culturing at a
reduced oxygen tension (5%) may be beneficial for progenitor cell
expansion. LaIuppa et al., Exp. Hematol., 26, 835-843 (1998);
McAdams, et al., Trends Biotechnol., 14, 341-359 (1996); Koller et
al., Ann. New York Acad. Sci., 665, 105-116 (1992); Koller et al.,
Exp. Hematol., 20, 264-270 (1992); Koller et al., Blood, 80,
403-411 (1992).
Suitable feeding protocols are known in the art or can be
determined empirically, and making such determinations is within
the skill in the art. In general, the higher the inoculum density
and the higher the cell density present in a culture, the more
often feeding of a hematopoietic culture (static or stirred) will
be required.
One protocol for feeding stirred or static cultures is replacing
50% of the culture medium every other day beginning on day 4 of
culture (referred to herein as "FP1"). FP1 results in very high
cell densities and has been found to give good levels of expansion
of total cells and CFC (progenitor cells) (see Examples 1 and 2).
However, FP1 has also been found to limit the expansion of total
cells, particularly post-progenitor cells (see Example 2).
As a consequence, another feeding protocol (referred to herein as
"FP2") was developed which provides for greater expansion of total
cells (see Example 2). In FP2, the cell density of the culture is
adjusted every other day or, preferably, every day to
1.5-2.0.times.10.sup.6 cells/ml. To make this adjustment of the
cell density, a portion of the spent culture medium containing
cells is replaced with fresh culture medium. The same effect can be
obtained by simply diluting the cell culture every other day or,
preferably, every day using fresh medium (with no removal of spent
medium or cells). To do so, a sufficiently large culture vessel or
multiple culture vessels (with the cell mixture proportionally
distributed between them) with adequate control of pH and oxygen
tension must be used to accomodate the greater volume. Further, the
total cell density could be measured on-line using a cell density
probe, and medium addition (and the removal of cells and spent
medium, if necessary) could be carried out continuously as part of
an automated system. Using FP2, expansion of total cells and CFC
(progenitor cells) is substantially increased compared to FP1 (see
Example 2, Table 1). Also, expansion of post-progenitor cells was
not inhibited as it was with FP1.
It should be noted that the combination of certain feeding
protocols and growth factors can stimulate a culture to contain a
large percentage of monocytes, and monocytes have a greater
q.sub.gluc and q.sub.lac than do other mature cells, although still
less than CFCs. If the percentage of monocytes is high enough, the
correlation of maximum % CFC with maximum q.sub.gluc and q.sub.lac
does not hold. Accordingly, feeding protocols which stimulate the
generation of large percentages of metabolically active monocytes
should be avoided if it is desired to rely on glucose consumption
or lactate production for determination of CFC content of a
hematopoietic cell culture. This is not a major limitation,
however, because large numbers of monocytes are not normally
desired for transplantation therapies. Moreover, it appears that
monocytes do not consume large amounts of oxygen, so the
correlation between % CFC and q.sub.02 can be used to estimate CFC
content in cultures containing monocytes (see Example 2).
To employ the methods of the invention, the reference and
experimental cultures must be cultured under "essentially the same"
conditions. "Essentially the same" means: (1) that the source of
the cells is the same for the reference and experimental cultures;
(2) that the cells in the reference and experimental cultures are
from similar samples; (3) that cells are processed in the same
manner; (4) that the inoculum densities of the reference and
experimental cultures are similar; (5) that the same cytokines are
used in the reference and experimental cultures; (6) the same
feeding protocol is used; and (7) that the initial physicochemical
conditions of the reference and experimental cultures are the same.
Quantitative measurements (e.g., many of the physicochemical
conditions) are normally subject to measurement errors. Thus,
"same" means that such quantitative measurements are the same
subject to normal measurement errors. As noted above, the inoculum
densities need only be "similar." For instance, if mononuclear
cells are used, the CD34.sup.+ cell contents of the reference and
experimental cultures may differ by about .+-.1% (by contrast, the
normal measurement error for CD34.sup.+ cells is about .+-.0.1%).
For example, if the experimental culture contains 5% CD34.sup.+
cells, the reference culture should contain 4-6% CD34.sup.+
cells.
The reference culture can be performed at any time prior to
performing the experimental culture. However, the reference culture
is preferably repeated whenever a reagent or equipment is changed
(e.g., a new supplier of a reagent, a new batch or lot of a
particular reagent, a new incubator).
Also, several reference cultures can be performed covering normal
operating conditions. For instance, reference cultures can be
performed employing the different culture conditions usually used
in a particular laboratory or the different culture conditions
usually used in a culture intended for a particular end use of the
cultured cells. Also, several reference cultures can be performed
employing the usual range of CD34.sup.+ cells encountered in a
particular source of cells.
The invention provides a method of determining the content of
progenitor cells in a hematopoietic cell culture. As used herein,
"content" refers to the relative number of progenitor cells in the
culture and to the absolute number of progenitor cells in the
culture. For instance, % CFC or total CFC in a hematopoietic cell
culture can be determined. Also, the time at which progenitor cells
are at a maximum concentration can be determined.
The content of progenitor cells in a hematopoietic cell culture is
determined by measuring one or more metabolic parameters.
"Metabolic parameter" is used herein to mean any measure of the
metabolic activity of the cells. "Metabolic parameters" which can
be used to determine the progenitor cell content of hematopoietic
cell cultures include glucose consumption, lactate production,
oxygen consumption, pyruvate consumption, ammonia production,
carbon dioxide production, and production or consumption of one or
more amino acids.
Devices and methods for measuring glucose consumption, oxygen
consumption, and lactate production of cultures are well known.
Suitable devices for making such measurements are available
commercially from, e.g., Yellow Springs Instruments, Kodak, Ingold
and Instrumentation Laboratories. See also Examples 1 and 2.
Devices and methods for measuring pyruvate consumption, ammonia
production, carbon dioxide production, and production or
consumption of amino acids are also well known. For instance, amino
acid consumption or production can be measured by high pressure
liquid chromatography (devices available commercially from Waters
or Hewlett-Packard), ammonia production can be measured using an
ion-specifc electrode (available commercially from Orion), pyruvate
consumption can be measured using an enzymatic assay (device
available from Sigma), and carbon dioxide production can be
measured using a blood gas analyzer (devices available from
Instrumentation Laboratories or Corning).
The one or more metabolic parameters must be measured in the
reference culture at two or more selected times. Preferably five
measurements at five selected times, more preferably eight
measurements at eight selected times, will be made. Even more
preferably, measurements will be made once a day every day for ten
or more days. Most preferably, measurements will be made
continuously.
The one or more metabolic parameters must also be measured at one
or more
selected times for the experimental culture. At least one
measurement should be made early in the culture (preferably at
24-48 hours) to confirm that the culture is performing as expected
based on the performance of the reference culture. If not,
corrective action can be taken or another source of cells for the
end use can be found. Several additional measurements should be
made during the experimental culture to confirm that the maximum or
desired number of progenitor cells is present in the culture prior
to harvesting. In a preferred embodiment, the one or more metabolic
parameters is(are) monitored daily. Even more preferably, the
experimental culture is monitored continuously, and the
measurements are fed to a computer which calculates the content of
the progenitor cells. With daily or continuous monitoring, the best
time to harvest the culture can be easily identified. Also, any
problems with the culture can be identified at the earliest
possible moment.
The number of progenitor cells (CFC) in the reference culture must
be determined. Assays for CFC (as a group or specific types) are
well known in the art (see the Background section and Example 1).
The total number of nucleated cells used to initiate the CFC assay
is also determined by methods well known in the art, preferably
using a Coulter Counter or other similar apparatus (see Example 1).
Using these two measurements, the % CFC can be calculated at each
selected time.
The density or concentration of nucleated cells in the reference
culture is also measured at each of the selected times. Suitable
methods of doing so are the same as for measuring nucleated cells
for the CFC assays. For instance, the density of nucleated cells
can be measured using a Coulter Counter or other similar apparatus
(see Example 1). When continuous measurements are employed, cell
density or concentration can measured using an optical probe.
Suitable optical probes are available from Aquasent, Ingold,
Monitek and Wedgewood.
Finally, the measurements of the one or more metabolic parameters
for the experimental culture are compared to those measured for the
reference culture to determine the content of progenitor cells in
the experimental culture at one or more selected times. The
comparison can be made graphically or by use of mathematical
equations setting forth the relationship between the measured
metabolic parameter(s) and the content of progenitor cells.
For example, the total glucose consumed, total oxygen consumed, or
the total lactate produced per ml by a reference culture is
measured once a day every day for ten days. The density of
nucleated cells and % CFC of the reference culture are also
determined at each of these times. The measured cell density, %
CFC, and glucose consumption, oxygen consumption or lactate
production of the reference culture are used to calculate
q.sub.gluc, q.sub.02 or q.sub.lac, and to determine
.alpha..sub.gluc and .beta..sub.gluc, .alpha..sub.02 and
.beta..sub.02 or .alpha..sub.lac and .beta..sub.lac using the
equations set forth in Examples 1 and 2 below. The values of
.alpha..sub.gluc and .beta..sub.gluc, .alpha..sub.02 and
.beta..sub.02, or .alpha..sub.lac and .beta..sub.lac can be
determined as described in Examples 1 and 2 employing a graph of
q.sub.gluc, q.sub.02 or q.sub.lac versus % CFC. Alternatively, the
values of .alpha..sub.gluc and .beta..sub.gluc, .alpha..sub.02 and
.beta..sub.02 or .alpha..sub.lac and .beta..sub.lac can be
determined using linear regression analysis employing least squares
fit, preferably by computer, to calculate the values of
.alpha..sub.gluc and .beta..sub.gluc, .alpha..sub.02 and
.beta..sub.02 or .alpha..sub.lac and .beta..sub.lac from equation
(9), (21) or (6), respectively.
Then, the density of nucleated cells of, and total glucose
consumed, total oxygen consumed or the total lactate produced per
ml by, an experimental culture are measured. These measurements and
the values of .alpha..sub.gluc and.beta..sub.gluc, .alpha..sub.02
and .beta..sub.02 or .alpha..sub.lac and .beta..sub.lac for the
reference culture are used to calculate the % CFC in the
experimental culture at the one or more selected times using
equation (9), (21) or (6), respectively, set forth in Examples 1
and 2 below. As described in Examples 1 and 2, the % CFC reaches a
maximum and then declines. If it is desired to harvest the
experimental culture at the time of maximum % CFC, measurements of
the glucose consumption, oxygen consumption, or lactate production
should begin early in culture and continue until the maximum % CFC
is achieved. Of course, as noted above, continuous monitoring of
the experimental culture will give the time of maximum % CFC most
precisely.
Alternatively, the measured glucose consumption, oxygen consumption
or lactate production of the experimental culture is used to
calculate Q.sub.gluc, Q.sub.02 or Q.sub.lac, the concentration of
nucleated cells (X.sub.i) in the experimental culture is measured,
and the measured X.sub.i and the calculated Q.sub.gluc, Q.sub.02 or
Q.sub.lac for the experimental culture and the values of
.alpha..sub.gluc and.beta..sub.gluc, .alpha..sub.02 and
.beta..sub.02 or .alpha..sub.lac and .beta..sub.lac for the
reference culture are used to calculate the total number of CFC at
the selected time(s) for the experimental culture. These
calculations are made using equations (7), (22) or (4),
respectively, set forth in Examples 1 and 2. As with % CFC, if it
is desired to harvest the experimental culture at the time of
maximum total CFC, measurements of the glucose consumption, oxygen
consumption or lactate production should begin early in culture and
continue until the maximum total CFC is achieved. Of course,
continuous monitoring of the experimental culture will give the
time of maximum total CFC most precisely.
Monitoring glucose consumption, oxygen consumption or lactate
production is also useful in determining when cells have exited
from quiescence. This exit is evidenced by a rapid increase in
q.sub.gluc, q.sub.02, or q.sub.lac, indicating the beginning of
rapid proliferation. This information could be useful for
determining when to introduce genetic material to hematopoietic
cells for the purpose of gene therapy, since most gene therapy
transfection protocols require that the host cells be in a cycling
state.
EXAMPLES
Example 1
In this example, the correlation of glucose consumption and lactate
production with colony-forming cell content in hematopoietic cell
cultures was investigated.
A. Media and Reagents
Media--Human long term medium (HLTM), which was used as the
standard serum-containing medium, consists of McCoy's 5A basal
medium (Sigma, St. Louis, Mo.), 12.5% heat inactivated horse serum
(Sigma), 12.5% fetal bovine serum (Hyclone, Logan, Utah), 1 mM
sodium pyruvate (Sigma), 1% MEM vitamin solution (Irvine
Scientific, Irvine, Calif.), 1% MEM amino acid solution (Irvine
Scientific), 1% MEM non-essential amino acid solution (Irvine
Scientific), 10.sup.-4 M monothioglycerol (Sigma), 2 mM L-glutamine
(Sigma), and 50 .mu.g/ml gentamycin sulfate (Gibco, Grand Island,
N.Y.). XVIVO-20 (BioWhittaker, Walkersville, M.d.), was used as the
standard serum-deprived medium.
Cytokines--All cytokines used were purified recombinant human
factors. Interleukin-3 (IL-3, Sandoz, East Hanover, N.J.) was used
at 5 ng/ml, IL-6 (Sandoz) at 10 ng/ml, stem cell factor (SCF,
Amgen, Thousand Oaks, Calif.) at 50 ng/ml, Flt3 ligand (Flt3-l,
Immunex, Seattle, WA) at 50 ng/ml, granulocyte-colony stimulating
factor (G-CSF, Amgen) at 1.5 ng/ml, granulocyte-macrophage-CSF
(GM-CSF, Immunex) at 2 ng/ml, and erythropoietin (Epo, Amgen) at 28
ng/ml in liquid cultures.
B. Cells and Cell Separation Procedures
Patient samples (Response Oncology; Memphis, Tenn.) of peripheral
blood (PB) were collected after informed consent under protocols
approved by the respective Institutional Review Boards. Apheresis
products were collected from cancer patients following stem cell
mobilization regimens consisting of treatment with G-CSF with or
without chemotherapy. The samples were used as received; density
gradient separation of the mononuclear cell (MNC) fraction was not
required due to minimal erythrocyte content. Umbilical cord blood
(CB) samples were provided by Northwestern University Memorial
Hospital (Chicago, Ill.). CB MNC were isolated from the whole
sample by density gradient separation on Histopaque (1.077 g/ml,
Sigma). Positive selection of CD34 antigen-bearing cells (CFC are
contained within the CD34.sup.+ cell population) was accomplished
by utilizing MiniMACS (Miltenyi Biotech, Inc., Sunnyvale, Calif.)
magnetic separation columns following the directions of the
manufacturer. The number of nucleated cells was determined on a
Coulter Counter Multisizer (Coulter Electronics, Hialeah, Fla.)
after treatment with cetrimide solution (90 g cetrimide
(hexadecyltrimethylammonium bromide)powder (Sigma), 25 g NaCl
(Sigma), 1.1 g. EDTA (Sigma) in 3 L deionized water) to lyse the
cells and release the nuclei. The error associated with the
preparation and measurement of cell density was estimated to be
.+-.5%.
C. Methylcellulose Colony Assays
The numbers of granulocyte, monocyte/macrophage, erythroid, and
mixed-lineage progenitor cells were determined using a
methylcellulose colony assay as described previously (Koller et
al., Blood, 80, 403-411 (1992)), with slight modifications. The
1.1% methylcellulose medium was supplemented with IL-3, IL-6, SCF,
GM-CSF, G-CSF, and Epo at the concentrations listed above for
liquid culture, with the exception of Epo, which was added at a
concentration of 83 ng/ml. Cultures were plated at seeding
densities ranging from 2,000 cells/ml to 15,000 cells/ml for fresh
and cultured MNC and from 500 cells/ml to 10,000 cells/ml for
cultures initiated with CD34.sup.+ cells. The inoculum density for
methylcellulose culture was determined both by the degree of total
cell expansion in the culture and the day of culture since the
cloning efficiency drops as the total cell expansion rises. The
methylcellulose plating density was, therefore, increased as the
culture expanded in an effort to maintain a total of 100-300 CFC
per dish. By so doing, the effects of either overplating or
underplating the methylcellulose culture that occur when a fixed
seeding density is used at all time points are avoided. The
methylcellulose cultures were incubated for 14 days in a humidified
atmosphere of 5% O.sub.2 and 5% CO.sub.2 (balance N.sub.2). At the
end of the incubation period, colonies of 50 or more cells were
enumerated as either CFU-GM (including CFU-G and CFU-M), BFU-E, or
CFU-Mix through inspection on a dark field stereomicroscope (Zeiss,
Batavia, Ill.). The majority of the cultures did not contain
detectable numbers of CFU-Mix, so that CFU-Mix was neglected in the
analysis. The error associated with enumerating CFC was estimated
using Poisson statistics. From the number of each CFC type counted
and the plating density, the percentage of each CFC type or total
CFC present in the sample was calculated as ##EQU2## D. Stirred
Hematopoietic Culture
Stirred cultures were carried out in 100-ml spinner flasks (Bellco
model 1967 with model 1965 agitator assembly) with an agitation
rate of 30 RPM (see Collins et al., Chemical Engineering Progress,
Supplement 1, 57a (1996) and Collins et al., Biotechnol. Bioeng.,
59, 534-543 (1998), the disclosures of which are incorporated
herein by reference). The spinners were not siliconized prior to
use, as this was found to be unnecessary. The cultures were fed
every 2 days, beginning at day 4, by pipette removal of one half of
the cell suspension, centrifugation at 300 g for 10 minutes,
removal of the spent medium, and return of the cells with fresh
equilibrated medium to the spinner flask (this feeding protocol is
referred to herein as FP1). The spinner flasks were maintained
within a humidified incubator at 5% O.sub.2 and 5% CO.sub.2
(balance N.sub.2). Samples containing cells and medium were removed
by pipette, and the number of nucleated cells enumerated using a
Coulter counter. Medium supernatant samples were frozen at
-20.degree. C. and retained for metabolite analysis.
E. Static Hematopoietic Culture
Static cultures were carried out in T-75 flasks (Falcon, Lincoln
Park, N.J.) for MNC culture or T-25 flasks (Falcon) for CD34.sup.+
cell cultures. Static cultures were maintained in the same manner
as stirred cultures.
F. Metabolic Assays and Calculations
Medium supernatant samples were thawed and subsequently centrifuged
for 10 minutes at 14,000 RPM using an Eppendorf model 5415C
centrifuge to remove any particulates that could potentially foul
the membranes of the YSI model 2700 glucose/lactate analyzer
(Yellow Springs Instruments, Yellow Springs, Ohio). The analyzer
was calibrated after every six samples to enhance the accuracy of
the assays. The manufacturer's stated assay precision is .+-.2%,
while the linearity is .+-.2% between 0-13.9 mmole and .+-.5%
between 13.9-138.9 mmole for glucose and .+-.2% between 0-5.6 mmole
and .+-.5% between 5.6-29.1 mmole for lactate.
Volumetric glucose consumption and lactate generation rates (Q,
.mu.mole/ml/hr) were calculated using the second-order central
slope method, as follows. The total glucose consumed or lactate
generated was calculated for each time point. For any time point
(t.sub.i), the first order forward (f, from time t.sub.i to
t.sub.i+1) and backward (b, from time t.sub.i-1 to t.sub.i) slopes
of the total consumption or generation curve were calculated by
dividing the point-to-point metabolite consumption or generation
differences by the point-to-point differences in time. The
volumetric rate at time t.sub.i was then calculated by taking a
time-distance weighted average of these two slopes: ##EQU3##
Specific metabolic rates (q, pmole/cell/hr) at any time t.sub.i
were obtained by dividing the volumetric rate by the nucleated cell
density X.sub.i (cells/ml) at time t.sub.i : ##EQU4## The nucleated
cell density does not account for enucleated red blood cells (RBC,
final stage erythroid cells do not have a nucleus), which may be
present in the culture. This enucleated population was not
accounted for because the small number of RBC present at
inoculation are typically no longer detectable by day 3 and the
formation of enucleated RBC was not generally observed in the
cultures (as determined by phenotypic examination).
G. Correlation of Culture CFC Content with Specific Metabolic
Rates
In examining the specific glucose consumption rate (q.sub.gluc) and
lactate generation rate (q.sub.lac) for cultures carried out in
spinner flasks, it was observed that both q.sub.gluc and q.sub.lac
increased from time zero until a maximum was attained. After that
time, both q.sub.gluc and q.sub.lac fell until they reached a
minimum value that was maintained until the end of the culture. A
similar decrease (after reaching a maximum) in the fraction of
cells in a culture that were CFC (% CFC) suggested a relationship
between % CFC and q.sub.gluc (or q.sub.lac). When q.sub.gluc,
q.sub.lac, and % CFC were plotted vs. time on the same graph, they
rose and fell simultaneously (FIG. 1), with the maximum q.sub.gluc
or q.sub.lac observed when the percentage of CFC in culture was the
greatest. This suggested that rapidly proliferating CFC have a much
greater metabolic demand than more mature cells.
In all experiments conducted, similar trends were observed for
glucose consumption and lactate production rates. Lactate data
generally displayed less scatter than did glucose data. This was
especially true early in cultures inoculated at low cell densities
where the error in the glucose assay was of the same order of
magnitude as the amount of glucose consumed. For simplicity, most
of the remaining discussion is limited to lactate production
rates.
The data shown in FIG. 1 are for a CB MNC spinner flask culture
conducted using XVIVO-20 with IL-3, IL-6, SCF, G-CSF, GM-CSF, and
Epo at an inoculum density (ID) of 3.85.times.10.sup.5 cells/ml.
However, the relationship between % CFC and q.sub.lac is not
dependent upon a particular cell source, culture system, medium
type, ID, or cytokine combination. To date, this relationship has
been observed in more than ten CB MNC, 47 PB MNC, and 18 PB
CD34.sup.+ cell cultures carried out under a variety of conditions.
For example, the correlation between q.sub.lac and % CFC shown in
FIG. 1 was also evident in a parallel T-flask culture (FIG. 2). For
a
given cytokine combination, cultures initiated at different ID from
the same PB MNC sample exhibited the same interdependence of
q.sub.lac with % CFC (FIG. 3). The coincidence of q.sub.lac and %
CFC was also maintained for different cytokine combinations in
cultures initiated at the same ID from the same PB MNC sample
(FIGS. 4 and 5). The correlation was evident in both
serum-containing (FIGS. 3, 6, and results not shown) and serum-free
(FIGS. 1, 2, 4, 5, and results not shown) culture media. The shape
of the q.sub.lac profile is not always the same from culture to
culture, but the time of maximum q.sub.lac still corresponds to the
time of maximum CFC content. The coincidence of maxima in q.sub.lac
and % CFC also extended to cultures that exhibited local minima and
maxima (FIGS. 3 and 4B), although the majority of the cultures did
not realize increases in % CFC once the decline from maximum % CFC
had begun.
The maximum CFC content observed in MNC cultures is typically on
the order of 10%. The coincidence of maxima for q.sub.lac and % CFC
held true in cultures with higher CFC content, such as those
inoculated with CD34.sup.+ cells. In CD34.sup.+ cell cultures, the
% CFC achieved was much greater (as high as 40%), and the time to
maximum % CFC was shorter than in MNC cultures. Despite the
differences in initial culture population, maximum CFC content
attained, and growth kinetics, the % CFC content was reasonably
paralleled by q.sub.lac in CD34.sup.+ cell cultures with three
different ID (FIG. 6). However, it should be noted that the point
corresponding to the first calculated q.sub.lac value for the three
cultures shown in FIG. 6, as well as for most other CD34.sup.+ cell
cultures, was well below that expected for a culture with such a
high % CFC. The low q.sub.lac can be attributed to the fact that
primitive hematopoietic cells are predominantly in quiescence at
the onset of culture (Traycoff et al., Exp Hematol, 22, 1264-1272
(1994); Gore et al., Exp Hematol, 23, 413-421 (1995)), and would
therefore not be expected to demonstrate the high rate of lactate
generation associated with rapid proliferation. This period of
quiescence applies to MNC cultures as well, but the % CFC present
at the beginning of a MNC culture is much lower than that for a
CD34.sup.+ cell culture.
H. Modeling Cell Metabolism
An exact model of total lactate production would consider each
distinct cell type, such that: ##EQU5## where i represents each
individual cell type and n.sub.i, the number of cells of type i.
However, the large number of hematopoietic cell types makes this
model unwieldy. The dramatic decrease in q.sub.lac during the
differentiation from CFC to post-progenitor cells (for example, see
FIG. 5A at 120 hours) suggests that a two-population (CFC and other
cells) model provides an adequate description of lactate
production, such that:
where .alpha..sub.lac is the q.sub.lac value for a CFC,
.beta..sub.lac is the q.sub.lac value for a non-CFC, X.sub.i is the
concentration of total nucleated cells in the culture, and [CFC] is
the concentration of total CFC in the culture. Equation (4) can be
normalized by dividing both sides by X.sub.i. Upon rearrangement,
the following relationship is obtained: ##EQU6## or ##EQU7##
If this two-population model adequately describes the data, a plot
of q.sub.lac versus the % CFC in a culture will yield a straight
line with the y-intercept giving q.sub.lac for a non-CFC and the
slope yielding the difference between q.sub.lac for a CFC and that
for a non-CFC.
FIG. 7 shows the linear relationship between q.sub.lac and % CFC
for a CB MNC culture in HLTM with IL-3, IL-6, SCF, G-CSF, GM-CSF,
and Epo. The line though the data points was generated by linear
regression. In FIG. 7, the y-intercept .beta..sub.lac (q.sub.lac
for a non-CFC) is 2.5.times.10.sup.-8 .mu.mole/cell/hr. The
calculated .alpha..sub.lac for the data in FIG. 7 is
5.3.times.10.sup.-6 .mu.mole/cell/hr. Since .alpha..sub.lac is
approximately 200-fold greater than .beta..sub.lac, this
substantiates the hypothesis that the average CFC has a much
greater q.sub.lac than does a more differentiated cell.
Regression analysis (FIG. 8) was also performed on the data
previously presented in time-course form in FIG. 3. As before, a
reasonable straight-line relationship was obtained between
q.sub.lac and % CFC. Again, the calculated .alpha..sub.lac is much
greater than .beta..sub.lac for each plot. The degree of
correlation varied for the different cultures, but, in most cases,
points that deviate from the regression line are explained by the
errors associated with the % CFC and q.sub.lac calculations.
As mentioned above, hematopoietic progenitors are typically
quiescent at the onset of culture. Quiescent CFC would not be
expected to exhibit high lactate production rates. Thus, the
proposed model would be expected to only describe lactate
production for cells that have exited quiescence. This is not a
major limitation because cells typically leave quiescence within 24
to 48 hours of culture (Traycoff et al., Exp Hematol, 22, 1264-1272
(1994)). Indeed, linear regressions of the FIG. 6 data for
CD34.sup.+ cell cultures describe the relationship between % CFC
and q.sub.lac reasonably well after the first time point in the
culture, which is associated with the lag phase, is removed (FIG.
9).
The proposed model could be used to predict the total CFC content
if all cultures had the same values for .alpha..sub.lac and
.beta..sub.lac. However, these parameters vary with culture
conditions such as cytokine combination, inoculum density, and cell
type. It is more likely that the same .alpha..sub.lac and
.beta..sub.lac values will be obtained for samples with similar
CD34.sup.+ cell content cultured under identical conditions.
Indeed, when three different PB MNC samples with similar day zero
CD34.sup.+ cell content were cultured under identical conditions,
the values for .alpha..sub.lac and .beta..sub.lac were similar
enough that the data could be pooled into one correlation, as shown
in FIG. 10. The data demonstrate that changes in the CFC content of
a culture can be followed by monitoring q.sub.lac. A correlation
between q.sub.lac and % CFC was evidenced over a wide variety of
conditions, including spinner flask (FIGS. 1, 3, 4, 5) and T-flask
(FIGS. 2, 6) cultures with MNC (FIGS. 1, 2, 3, 4, 5) and CD34.sup.+
cells (FIG. 6) in serum-containing (FIGS. 3, 6) and serum-free
media (FIGS. 1, 2, 4, 5) with different cytokines (FIGS. 4, 5) and
inoculum densities (FIGS. 3, 6).
The coincidence of maximum % CFC with maximum q.sub.lac may be
useful in deciding when to manipulate a hematopoietic culture. For
example, for some applications of gene therapy using ex vivo
expanded cells, it might be best to initiate gene transfer when the
progenitor cell content is highest (i.e., the time of maximum % CFC
and q.sub.lac). However, if the expanded cells are to be used
directly for transplantation, it will likely be more beneficial to
harvest when the total content of CFC reaches a maximum. Any
particular clinical protocol is likely to be restricted to a single
cell type (PB or CB; MNC or CD34.sup.+ cells), culture system,
cytokine combination, and inoculum density. The correlation shown
in FIG. 10 indicates that it will be possible to identify unique
.alpha..sub.lac and .beta..sub.lac values for these cultures, or at
least distinct .alpha..sub.lac and .beta..sub.lac values for
different ranges of CD34.sup.+ cell content in MNC cultures. Since
the CD34.sup.+ cell content is routinely measured for hematopoietic
cell harvests, this means that the appropriate .alpha..sub.lac and
.beta..sub.lac values would be available at the beginning of each
culture. In this event, it should be possible to determine the
total CFC content (from equation 4), as well as the % CFC (from
equation 6), at any time point in the culture, thereby realizing
real-time determination of CFC content in culture.
A two-population model also provides an adequate description of
glucose consumption. See FIGS. 11A-B and 12A-B. Equations for the
two-population model for glucose are:
Example 2
The successful application of spinner flask culture for
hematopoietic cells from a variety of sources in both
serum-containing and serum-free media was described in Example 1.
Spinner flask systems have also been used to culture bone marrow
mononuclear cells (BM MNC) (Zandstra et al., Bio/Technology, 12,
909-914 (1994); Sardonini and Wu, Biotechnol. Prog., 9, 131-137
(1993)).
A well-controlled, closed, and reproducible culture environment,
such as that offered by stirred bioreactors, will undoubtedly prove
advantageous for clinical applications, especially considering the
scale involved for clinical cultures. The culture volume employed
for recent clinical trials averaged about 5 liters (Zimmerman et
al., J. Hematotherapy, 4, 527-529 (1995); Williams et al., Blood,
87, 1687-1691 (1996)). Fifty (50) T-150 flasks each containing 100
ml culture medium or 20 gas-permeable 300 cm.sup.2 culture bags
each containing 250 ml would be necessary to accommodate this
volume. These phase I clinical trials were conducted to determine
the safety of infusing expanded cells. As trials continue, greater
numbers of cells will undoubtedly be transfused in an effort to
increase the efficacy of expansion protocols. Although peripheral
blood (PB) MNC-derived natural killer (NK) cells have been cultured
in a stirred bioreactor (Pierson et al., J. Hematotherapy, 5,
475-483 (1996)), controlled, stirred-tank bioreactor systems have
not yet been reported for the culture and characterization of
myeloid-lineage hematopoietic cells.
In this example, the effects of different hematopoietic populations
(CFC and more mature cells) on oxygen consumption, glucose
consumption, and lactate production, and on the ratio of glycolytic
to oxidative metabolism were examined. Also, culturing
hematopoietic cells in bioreactors was investigated.
A. Materials and Methods
Medium. HLTM (see Example 1) was used as the culture medium. HLTM
was supplemented with purified recombinant human cytokines: 5 ng/ml
IL-3 (Novartis, East Hanover, N.J.), 50 ng/ml IL-6 (Novartis), 50
ng/ml SCF (Amgen), 1.5 ng/ml G-CSF (Amgen), 2 ng/ml GM-CSF
(Immunex), and 28 ng/ml Epo (Amgen).
Cells and Cell Separation Procedures. Patient samples (Response
Oncology) of mobilized PB MNC were collected as described in
Example 1. Apheresis products were also collected as described in
Example 1. Samples in 15-ml polystyrene test tubes containing
anticoagulant citrate dextrose were stored and shipped under
ambient conditions and used as received within 2-3 days of
collection; enrichment of the MNC fraction was not required due to
minimal erythrocyte content. CB samples were provided by
Northwestern University Memorial Hospital (Chicago, Ill.).
Erythrocytes were depleted from the whole sample by ammonium
chloride lysis (Denning-Kendall et al., Exp. Hematol., 24:1394-1401
(1996)). All samples were incubated for 2-4 days at
1.2-2.times.10.sup.6 cells/ml in T-150 flasks (Falcon, Lincoln
Park, N.J.) prior to inoculation in either a T-75 flask or a
stirred bioreactor. This pre-incubation period, which may prove not
to be necessary, was designed to acclimate the cells to the culture
conditions of the experiment in the absence of potential
fluid-mechanical damage. In addition, it allowed a lag phase to be
avoided by transferring exponentially-growing cells into the
bioreactor. The number of nucleated cells was determined on a
Coulter Counter Multisizer after cetrimide treatment to lyse the
cells and release the nuclei.
Methylcellulose Colony Assays. The numbers of granulocyte and
monocyte/macrophage (collectively CFU-GM), and erthyroid (BFU-E)
progenitor cells (colony-forming cells, or CFC) were determined
using the methylcellulose colony assay described in Example 1.
Culture Conditions. Stirred bioreactor cultures were carried out in
a 400-ml B. Braun Biostat Q (B. Braun Biotech USA, Allentown, Pa.)
with an agitation rate of 30 rpm at a working volume of 150-200 ml.
The reactor's stainless steel agitator assembly was removed because
detrimental effects of stainless steel on hematopoietic cell
proliferation have been observed (LaIuppa et al., Journal of
Biomedical Materials Research, 36:347-359 (1997). A Bellco spinner
flask model 1965-250 agitator assembly was fitted into a
compression fitting on the headplate of the reactor. The inside
diameter of the Biostat Q is the same as that for the Bellco
spinner flask model 1967-100 used in Example 1. The agitation setup
was therefore identical (except for the longer agitator shaft) to
that employed (d.sub.i /D=0.8) in spinner flasks in Example 1. The
reactor was maintained within a 37.degree. C. incubator and was
fitted with dissolved oxygen (DO, Ingold, Wilmington, Mass.) and pH
(Ingold) probes, which were interfaced to a personal computer (PC)
via the Workbench PC program (Omega, Stamford, Conn.). DO was
controlled at 50% of air saturation through headspace addition of
humidified O.sub.2, N.sub.2, and air. pH was controlled at
7.33.+-.0.03 through headspace addition of humidified CO.sub.2.
Static cultures were carried out in T-75 flasks (Falcon, Lincoln
Park, N.J.) maintained at 37.degree. C. inside a 5% CO.sub.2
(balance air) incubator.
Feeding Protocols. The cultures in Experiment 1 were fed every 2
days, beginning at day 4, by pipette removal of one half of the
cell suspension, centrifugation at 300.times.g for 10 minutes,
removal of the spent medium, and return of the cells with fresh
equilibrated medium to the culture vessel, thereby maintaining a
constant culture volume. This feeding protocol was used in the
spinner-flask and T-flask culture systems for PB and CB MNC (see
Example 1) and is designated Feeding Protocol 1 (FP1). In
subsequent experiments, cultures were diluted daily (Experiments 3
and 4) or every two days (Experiment 2) to a density of
1.5-2.times.10.sup.6 cells/ml. Practically, 25-45% of the culture
broth (depending on the measured cell density) was removed and the
reactor replenished with fresh medium (Feeding Protocol 2, FP2).
T-flask controls were fed with a volumetric exchange equivalent to
that in the bioreactor. Cell expansion ratios were calculated by
determining the total cells that would have been produced in the
vessel, assuming that the removed cells expanded in an identical
manner as the remaining cells. Medium supernatant samples were
frozen at -20.degree. C. and retained for metabolite analysis.
Metabolic Assays and Calculations. Medium supernatant samples were
thawed and subsequently analyzed on a YSI model 2700
glucose/lactate analyzer as described in Example 1. Specific
glucose and lactate metabolic rates (q, umole/cell/hr) at any time
t.sub.i were calculated as follows: ##EQU9## X.sub.b and X.sub.f
are the log-mean average cell densities for the time periods before
and after time i. The log-mean cell density is the effective
average cell density. It takes into account the fact that cell
density increases exponentially with time. X.sub.b and X.sub.f are
calculated as follows: ##EQU10## where the superscripts "bd" and
"ad" represent the cell density before and after dilution,
respectively. Q.sub.b is the point-to-point volumetric glucose
consumption or lactate generation rater ##EQU11## from time
t.sub.i-1 to t.sub.i and Q.sub.f is point-to-point rate from time
t.sub.i to t.sub.i+1. For example, for lactate: ##EQU12## where the
superscripts "bd" and "ad" represent the lactate concentration
before and after dilution, respectively. Similar equations apply
for glucose consumption. The specific glucose and lactate metabolic
rates were calculated in this manner because of the abrupt changes
in cell density at each dilution step associated with FP2 (see FIG.
14) which did not occur with FP1.
The general relation describing the change in oxygen concentration
in the culture medium with time (t) is given by: ##EQU13## where
C.sub.02 is the oxygen concentration in the liquid
(.mu.mole/ml),
C.sub.02 * is the liquid oxygen concentration (.mu.mole/ml) in
equilibrium with the headspace gas, K.sub.L a (hr.sup.-1) is the
volumetric mass transfer coefficient, and Q.sub.02 is the
volumetric oxygen uptake rate (.mu.mole/ml/hr) by cells in the
culture. The Henry's coefficient for oxygen in water
(atm.multidot.1/mmole) is used to calculate C.sub.02 * from the
partial pressure of oxygen in the headspace and to calculate
C.sub.02 from the DO probe reading (the % of air saturation value
is first converted to oxygen partial pressure).
The specific oxygen uptake rate (q.sub.02, .mu.mole/cell/hr) was
determined for Experiments 1 and 2 using the steady-state method
described in Miller et al., Bioprocess Engineering 3:103-111
(1988), the complete disclosure of which is incorporated herein by
reference. Briefly, DO was measured using the DO probe. The
volumetric mass transfer coefficient, K.sub.L a, was experimentally
determined for cell-free medium by following the increase (or
decrease) in oxygen concentration when air (or nitrogen) was passed
through the vessel headspace. The oxygen concentration in the
vessel headspace was determined using the gas flow rates obtained
from calibrated rotameters. At steady-state, the derivative term in
equatio n (15) i s equal to zero, so that:
The specific oxygen uptake rate, q.sub.02 n is calculated from the
Q.sub.02 as follows:
where X.sub.i is the total cell concentration at the time that
Q.sub.02 is determined.
The specific oxygen uptake rate (q.sub.02, .mu.mole/cell/hr) was
determined for Experiments 3 and 4 using the dynamic method
described in Zhou and Hu, Biotechnol. Bloeng., 44:170-177 (1994),
the complete disclosure of which is incorporated herein by
reference, using a computer (PC) for data acquisition and control
and to perform calculations. Briefly, Q.sub.02 was first
determined. To do so, DO was increased to 65% of saturation with
air. Then, nitrogen gas was flushed into the culture vessel to
deplete the oxygen from the gas in the headspace, and DO was
allowed to decrease until reaching 30% of saturation. The time
profile of DO between 50% and 30% was used to calculate Q.sub.02
using equation (15). Equation (15) can be solved by integration as:
##EQU14## where C.sub.02 (t.sub.0) is the C.sub.02 at the beginning
of taking the reading (t.sub.0) and C.sub.02 (t.sub.f) is the
C.sub.02 at the end (t.sub.f). Since the headspace is swept with
nitrogen during the measurement, C.sub.02 * is set to zero in
equation (18). K.sub.L a was measured with cell-free medium before
the cultivation, as described above. The specific oxygen
consumption rate (q.sub.02) was calculated from the Q.sub.02 and
the total cell concentration (X.sub.i) using equation (17).
For Experiments 3 and 4, q.sub.02 was also calculated using the
steady-state method on a number of occasions. Good agreement
between the two calculation methods was noted.
The ratio of the lactate production rate to the oxygen consumption
rate (Y.sub.lac,ox) was calculated as: ##EQU15## B. Results
Effect of Culture System and Feeding Protocol on Total Cell and
CFU-GN Expansion. The first bioreactor experiment utilized FP1; the
total cell and CFU-GM expansion ratios for this experiment are
shown in FIGS. 13A-B. The bioreactor and T-flask had comparable
total cell expansion profiles until day 6, when the total cell
concentration in both cultures appeared to have plateaued. On day
6, in an effort to increase the culture growth rate, the bioreactor
cell density was diluted from 5.2.times.10.sup.6 cells/ml to
3.2.times.10.sup.6 cells/ml, while the T-flask cell density
(5.1.times.10.sup.6 cells/ml) was not changed. The bioreactor was
fed using FP1 both before and after this single dilution event.
After the dilution, the rate of total cell expansion in the
bioreactor increased again (as it did in the T-flask, although to a
lesser extent) and a separation was evident between the performance
of the two vessels.
It was hypothesized that a dilution-feeding protocol, such as FP2,
would increase the extent of total cell expansion in the cultures.
FP2 was, therefore, utilized for both the T-flask and the
bioreactor in subsequent experiments. FIG. 14 shows the cell
density in a CB MNC experiment (Experiment 3) utilizing FP2
throughout the culture period. Through repeated dilution and
feeding, cell growth was maintained for an extended period of time
in both the reactor and the T-flask. As suspected, FP2 resulted in
a much greater expansion of total cells in both vessels (FIGS.
15A-C) than did FP1 (FIG. 13A). If the expansion product is to be
infused into a patient for the purpose of re-constituting the
hematopoietic system following chemotherapy, CFU-GM cells are
likely to be particularly important, since they give rise to
critical infection-fighting granulocytes. CFU-GM expansion in
FP2-fed systems (FIGS. 16A-C) was much greater than that observed
in either the bioreactor or T-flask FP1-fed cultures (FIG.
13B).
Metabolic Rates in the Bioreactor. FIGS. 17A-D show specific
lactate (q.sub.lac) and oxygen (q.sub.02) metabolic rates, along
with culture CFC content, for the four bioreactor cultures. For the
FP1-fed culture (FIG. 17A), the time of maximum q.sub.lac
corresponds closely with the time of maximum % CFC and the
decreasing CFC content is paralleled by decreasing q.sub.lac
values. These observations are consistent with those from previous
experiments that utilized FP1 (see Example 1). In the current work,
a similar trend for q.sub.02 was observed, although the time of
maximum q.sub.02 was delayed slightly beyond that for
q.sub.lac.
However, the times of maximum q.sub.lac and % CFC did not always
correspond for the FP2 cultures. The first FP2-fed CB MNC culture
(FIG. 17B) did exhibit an initial correspondence (through 120
hours) between CFC content and q.sub.lac, although qiac did not
decrease to a low level and a significant secondary rise in
q.sub.lac was observed. The sole PB MNC culture (FIG. 17D)
exhibited good correspondence between q.sub.lac and % CFC. In
contrast, the q.sub.lac trend observed for the CB MNC culture shown
in FIG. 17C is quite different. Although an increase in q.sub.lac
after 150 hours is noted in FIGS. 17A, 17B, and possibly 17D, it
follows an earlier peak in q.sub.lac. In FIG. 17C q.sub.lac
increases steadily from the beginning of the culture and is still
increasing at the end of the culture period.
Interestingly, q.sub.02 shows a much smaller (if any) secondary
increase after the initial decline. As a result, the q.sub.02
profile more closely corresponds to the % CFC profile than does the
q.sub.lac profile for the FP2 cultures.
The q.sub.lac and q.sub.02 profiles in the bioreactor suggest two
things: (1) that cell types other than CFC may impact metabolic
rates, particularly q.sub.lac, and (2) that the extent to which
different metabolic pathways are utilized for energy generation
changes with cell differentiation. The ratio of lactate production
to oxygen consumption (Y.sub.lac,ox) was, therefore, examined.
Higher Y.sub.lac,ox values indicate a shift to glycolytic energy
production, while lower Y.sub.lac,ox values suggest increased
utilization of oxidative phosphorylation. FIGS. 18A-D show the
profiles for Y.sub.lac,ox, along with the % CFC for each culture.
In general, Y.sub.lac,ox dropped from a high value early in culture
and reached a minimum when the % CFC began to plateau at a low
value. Thereafter, Y.sub.lac,ox increased again, which suggests
that post-progenitor cells rely to a greater extent on glycolysis
for their energy needs. Indeed, in the one experiment analyzed
using flow cytometry (FIG. 18C), the increase in Y.sub.lac,ox (and
q.sub.lac, FIG. 17C) was paralleled by an increase in the total
percentage of CD11b.sup.+ and/or CD15.sup.+ cells. Post-progenitor
cells of the granulomonocytic (GM) lineage, such as developing and
mature monocytes and granulocytes, are included in the CD11b.sup.+
and CD15.sup.+ populations. It was observed that the yield of
lactate from glucose (q.sub.lac /q.sub.gluc) increased after 150
hours in culture (data not shown), further supporting the idea that
post-CFC more extensively utilize glycolysis for energy
production.
C. Discussion
It has previously been demonstrated that both CB and PB MNC can be
cultured in spinner flasks under a variety of conditions (see
Example 1). Here, it has been demonstrated that a bioreactor vessel
can be adapted to provide an acceptable culture environment for
hematopoietic cells. As far as is known, this is the first report
of a stirred bioreactor with pH and DO control being used for the
expansion of myeloid-lineage hematopoietic cells. This work,
therefore, represents the next step in the evolution of
hematopoietic culture from static systems to spinner culture to a
fully-instrumented, stirred bioreactor. Total cell expansion in the
bioreactor was similar to that for past experiments in spinner
flasks using the cell-retention (FP1) feeding protocol. However,
CFU-GM expansion for the sample used in Experiment 1 was less than
the .about.15-fold expansion typically observed in spinner flasks
for CB MNC using FP1 (not shown).
Frequent, dilution-type feeding (FP2) increased the expansion of
total cells and CFU-GM, relative to using the FP1 feeding protocol,
for CB MNC cultures (Table 1 and data not shown for FP1
spinner-flasks). Similarly, the maximum expansion of total cells
and CFU-GM for the one PB MNC FP2 bioreactor culture (Experiment 4)
were greater than those in PB MNC FP1 spinner-flask controls (Table
1). Dilution-type feeding protocols have been utilized for static
CD34.sup.+ cell cultures by many researchers (Moore and Hoskins,
Blood Cells, 20:468-79 (1994); Lill et al., Stem Cells, 12:626-637
(1994); Haylock et al., Blood, 80:1405-1412 (1992)), with total
cell expansion ratios that often exceed 1,000. In static cultures
of MNC and CD34.sup.+ cells, expansion ratios and the kinetics of
expansion are highly dependent on the cell inoculum density (ID)
(Koller et al., Biotechnol. Bioeng., 50:505-513 (1996); Haylock et
al., in: Hematopoietic stem cells: biology and therapeutic
applications (D. Levitt et al., eds., Marcel Dekker, Inc., New
York, pp. 491-517 (1995)). In general, lower ID leads to greater
total cell expansion and greater depletion of CFC. Higher ID
cultures achieve greater total cell and CFC numbers, but a lower
total cell expansion ratio, when compared to lower ID cultures.
Perhaps due to the high residual cell density (1.5-2.times.10.sup.6
cells/ml) in our FP2 cultures, increased total cell expansion did
not come at the expense of CFU-GM depletion. As shown in Table 1,
the maximum observed percentages of CFU-GM (cloning efficiency) in
the FP2 cultures were much greater than those observed for the FP1
bioreactor culture (CB MNC) or spinner-flask controls (PB MNC).
Increased total cell production with no decrease (or even an
increase) in CFC production suggests that feeding by dilution
released the cells from a "blocked" state--perhaps induced by
accumulation of endogenous cytokines or inhibitory metabolites in
FP1 cultures, which reach 6-12.times.10.sup.6 cells/ml. Although
the underlying mechanisms are not well understood, decreased
metabolic activity in high-density cultures has previously been
reported for cells of hematopoietic origin, and has been termed a
"crowding" effect (Sand et al., Blood, 50:337-346 (1977)).
Additional evidence for the concept of release from inhibition is
provided by differences in the lactate production pattern for FP2
cultures, with a secondary increase in q.sub.lac after the decline
in % CFC (see below).
For FP2, total cell (FIGS. 15A-C) and CFU-GM (FIGS. 16A-C)
expansion were similar in the bioreactor and T-flask cultures. In
contrast, using FP1 superior total cell and CFU-GM expansion was
observed in stirred (spinner) vessels vs. T-flasks for PB MNC
cultures inoculated with 1.2.times.10.sup.6 cells/ml (see Collins
et al., Biotechnol. Bioeng., 59:534-43 (1998), the complete
disclosure of which is incorporated herein by reference). BM MNC
cultures in perfused T-flasks benefit (in terms of total cell and
CFU-GM expansion) from a similar dilution-feeding protocol (Oh et
al., Biotechnol. Bioeng., 44:609-616 (1994)). In that perfused
T-flask system, repeated cell removal increased the available
culture surface area and alleviated mass transfer limitations.
Similarly, the enhanced mass transfer in stirred culture will
minimize gradients in DO and pH (by increased CO.sub.2 removal), as
well as inhibitory cytokines that may accumulate in the cultures.
The beneficial effect of increased mass transfer is less important
at lower cell densities, as evidenced by similar total cell and
CFU-GM expansion in spinner flasks vs. T-flasks using FP1 in PB MNC
cultures inoculated with 2.times.10.sup.5 cells/ml (see Collins et
al., Biotechnol. Bioeng., 59:534-43 (1998). Thus, by frequently
diluting the cell density in the FP2 experiments reported here, the
benefit of stirred culture on cell and CFU-GM production was
offset. This is consistent with the observation that the benefit of
increasing medium exchange rate in static BM MNC cultures is
greater at higher ID (Koller et al., Biotechnol. Bioeng.,
50:505-513 (1996). Together, these observations suggest that a
controlled cell density is indeed beneficial for the proliferation
of hematopoietic cultures.
It was previously demonstrated that CFC consume more glucose and
generate more lactate on a per cell basis than do more mature cells
(see Example 1). Here it is shown that oxygen consumption is also
higher for CFC (FIGS. 17A-D). The q.sub.02 values for our stirred
cultures ranged from 1.7.times.10.sup.-8 to 1.2.times.10.sup.-7
.mu.mole/cell/hr. The highest values are slightly lower than the
1-3.times.10.sup.-7 .mu.mole/cell/hr reported for hybridomas
(Wohlpart et al., Biotechnol. Bioeng., 37:1050-1053 (1991); McQueen
and Bailey, Biotechnol. Bioeng., 35:1067-1077, (1990); Glacken et
al., Biotechnol. Bioeng., 32:491-506 (1988); Miller et al., J.
Cell. Physiol., 132:524-530 (1987)), and are much higher than the
1-4.times.10.sup.-8 .mu.mole/cell/hr reported for murine bone
marrow cells (Lutton et al., Experientia, 28:850 (1972); Olander,
American Journal of Physiology, 222:45-48 (1972); Gesinski and
Morrison, Experientia, 24:296-297 (1968); Gesinski et al.,
Australian Journal of Biological Sciences, 21:1319-1324 (1968)),
human bone marrow cells (Peng and Palsson, Annals of Biomedical
Engineering, 24:373-381 (1996), and normal human granulocytes (Bird
et al., Cancer, 1009-1014 (September 1951)). The values reported
for murine cells were obtained in experiments without the growth
factor stimulation present in the present cultures. Since growth
factor stimulation increases glucose uptake in culture (Whetton et
al., EMBO J., 3:409-413 (1984); Whetton et al., J. Cell Sci.,
84:93-104 (1986); Spielholz et al., Blood, 85:973-980 (1995);
Hamilton et al., Biochem. Biophys. Res. Commun., 138:445-454
(1986)), it is likely that cytokine stimulation would also increase
q.sub.02. The values reported for human BM MNC culture, which are
much lower than those reported here, were obtained in
growth-factor-supplemented systems (Peng and Palsson, Annals of
Biomedical Engineering, 24:373-381 (1996). However, some of the
q.sub.02 values may have been obtained under oxygen-limited
conditions, which would decrease the value of q.sub.02 (Miller et
al., J. Cell. Physiol., 132:524-530 (1987); Sand et al., Blood, 50:
337-346 (1977)). Interestingly, the q.sub.02 value for normal human
granulocytes (2.2.times.10.sup.-8 .mu.mole/cell/hr; Bird et al.,
Cancer, 1009-1014 (September 1951)) is similar to the q.sub.02
values that observed here when CFC have been depleted.
For FP1 cultures, a correlation between q.sub.lac and % CFC is
almost always observed--with a close correspondence between the
maxima in q.sub.lac and % CFC and a low value for q.sub.lac after
the CFC are depleted (see Example 1). For most FP2 cultures, there
was still a maximum in q.sub.lac near the maximum in % CFC, but in
both FP2 CB MNC cultures there was a secondary increase in
q.sub.lac that exceeded the initial peak value. This suggests that
GM post-progenitors can also exhibit significant glycolytic
activity, and that the correlation between q.sub.lac and % CFC
developed for FP1 cultures is not fully valid for FP2 cultures.
However, this is not as great a limitation as it may seem because
q.sub.02 does not exhibit a significant secondary peak after the
peak in % CFC (FIGS. 17A-D). This suggests that GM post-progenitors
are highly glycolytic, but do not have high oxidative metabolic
activity. In this regard, the % of cells that stained positive for
CD15 and/or CD11b (GM markers) increased proportionally with
Y.sub.lac,ox in one CB MNC FP2 culture analyzed by flow cytometry
(FIG. 18C). Although flow cytometry analysis for the other
bioreactor experiments was not performed, similar increases in
Y.sub.lac,ox were observed. Both monocytes and granulocytes
developed concurrently in Experiment 3; therefore, determining
which cell type had the largest effect on Y.sub.lac,ox and
q.sub.lac is difficult. Previous work with monocytes (Cline, in
Formation and Destruction of Blood Cells, pages 222-239 (Greenwalt
and Jamieson eds., 1970)) has shown that these cells rely heavily
upon glycolysis for their energy needs. Cultures supplemented with
Flt3-1 and macrophage-colony stimulating factor (M-CSF) have been
demonstrated to produce nearly pure populations of monocytes
(Gabbianelli et al., Blood, 86, 1661-1670 (1995)). Under these
culture conditions (using FP1), steadily increasing q.sub.lac and %
CFC have been observed (see Example 1), trends similar to those
shown in FIG. 17C. Finally, other reports indicate that
granulocytes have very few mitochondria (Bainton et al., J. Exp.
Med., 134, 907-934 (1971)) and would, therefore, not consume much
oxygen.
The control of pH, DO, and other physicochemical parameters in a
stirred bioreactor will allow for more refined studies regarding
the effects of these parameters on hematopoietic cells. Production
of the large numbers of hematopoietic cells desired for clinical
applications may well benefit from feeding protocols that help
control the concentration of all species in a culture. Stirred
bioreactor systems are readily adaptable to perfusion feeding
protocols, which allow for extensive medium replacement while
maintaining a high cell density. The results indicate that
monitoring metabolic quotients for oxygen, lactate, and other
metabolites will allow practitioners to estimate the percentages of
CFC in culture.
TABLE 1 ______________________________________ Maximum observed
total cell expansion, CFU-GM expansion, and % CFU-GM for FP1 and
FP2 cultures. Maximum Total Cell Maximum Maximum Expansion CFU-GM
Expansion % CFU-GM ______________________________________ CB MNC
Experiment 1 (FP1) 9.9/6.4 1.6/4.7 1.1/1.4 Bioreactor/T-flask
Experiment 2 (FP2) 34.2/26 23.5/23.7 4.71/4.6 Bioreactor/T-flask
Experiment 3 (FP2) 201/207 29/33 3.2/3.05 Bioreactor/T-flask PB MNC
Control.sup.a 14.4 .+-. 6.1 6.8 .+-. 3.2 1.61 .+-. 0.33 Spinners
(FP1) Experiment 4 397 13.5 5.6 Bioreactor (FP2)
______________________________________ .sup.a The control value is
the average .+-. 1 SEM for 5 (n = 5) PB MNC spinner flask cultures
carried out in HLTM + IL3, IL6, SCF, GCSF, GMCSF, and Epo at an ID
of 2.15 .+-. 0.09 .times. 10.sup.5 cells/ml. The control were fed
using FP1 and represent the typical expansion seen in this type of
system.
D. Modeling Cell Metabolism
As demonstrated in Example 1, q.sub.lac and % CFC can be related
with a linear model when using FP1. The results obtained in this
example for q.sub.02 suggested that a similar relationship might be
obtained for q.sub.02 and % CFC for both FP1 and FP2. Utilizing the
same model as for q.sub.lac (see Example 1, section H), but
substituting q.sub.02 for q.sub.lac, the following equation was
obtained: ##EQU16## where, .alpha..sub.02 is the q.sub.02 for an
average CFC and .beta..sub.02 is the q.sub.02 for the average
non-CFC.
This model was tested on the data of the bioreactor experiments,
and the results are shown in FIGS. 19A-D. In FIGS. 19A-D, q.sub.02
is plotted versus % CFC. The q.sub.02 corresponding to the first
time point (.about.24 hours) for the plots of experiments 1 and 2
(FIGS. 19A and 19B) was eliminated from the model analysis. These
q.sub.02 values were much lower than expected for the corresponding
% CFC. This may reflect an incomplete exit from quiescence at the
time of this first measurement, which was a problem for the lactate
model when applied to CD34.sup.+ cells. The q.sub.02 corresponding
to the third time point (.about.96 hours) for the plot of
Experiment 3 (FIG. 19C) was also eliminated from the model
analysis. This point appears to be an outliner as compared to the
remaining data and as compared to all other experiments' q.sub.02
profiles. The straight lines shown in FIGS. 19A-D were generated by
linear regression. In these graphs, the y-intercepts correspond to
.beta..sub.02 (the q.sub.02 for the average non-CFC), and the
slopes of the lines yield the differences between q.sub.02 for a
CFC and that for a non-CFC.
As can be seen, the model described oxygen consumption well for the
data that were fit for both FP1 and FP2. It is expected, therefore,
that the oxygen model will prove to be reliable for determining %
CFC over a wide range of culture conditions.
Also, Q.sub.02 and total CFC can be related by the following
equation:
* * * * *